The consumption of plastic packaging has been increasing continuously for years. The demand for packaging that protects products, increases their shelf life and at the same time encourages consumers to buy those products with an appealing feel and look conflicts with the need for the lowest possible use of materials and an advantageous mass ratio of packaging to product. Plastic foils precisely perform this balancing act. An annual consumption of around 69,000 t, representing around 23% of the annual volume of plastic packaging waste generated in Austria, reflects that both consumers and markets have recognised the benefits of these materials. However, the material properties that make the use of plastic films attractive as packaging make sorting using near-infrared spectroscopy, which represents the state of the art in waste management, more difficult. As a result, the materially recyclable monolayer fraction together with the multilayer fraction are sent for downcycling at best and energy recovery at worst, which negatively affects the recycling rate and at the same time wastes valuable resources. This doctoral thesis contributes to facilitating the sorting of the film fraction. For this purpose, an inventory of the current situation in Austria is carried out in this work using a manual sorting of the light packaging fraction and a state-of-the-art analysis with a focus on near-infrared technology in waste management. Subsequently, in a comprehensive life cycle analysis, the effect of increased mechanical recycling of the film fraction is compared with thermal recycling and it is ascertained that up to 63% of greenhouse gas emissions could be prevented through improved recycling. Based on these findings, methods for the adaptation of near-infrared sorters are presented, which circumvent the properties of the film fraction that hinder near-infrared sorting by measuring in transflection. These methods allow material-based sorting of the film fraction on existing units. Furthermore, data analysis methods based on spectral decomposition are presented, which allow to further optimize the spectral quality, which has been improved by the hardware adaptation. Subsequently, machine learning algorithms are examined for their applicability for sorting mono- and multilayer films, whereby the support vector machine and a shallow neural network turned out to be the most suitable methods for a material-independent classification of film waste into single- and multilayer films.