By using spectral information beyond the visible range, the project explores whether dust-affected scenes can be analysed more robustly than with conventional camera systems. This is particularly relevant for agricultural machinery, where dust generated during field operations can strongly reduce visibility and impair the performance of perception systems. For this purpose, field and experimental data are collected under relevant agricultural conditions. These data are used to specify requirements for hyperspectral perception systems in dusty environments, develop and evaluate data processing methods, and benchmark dust reduction as well as object detection for hyperspectral imaging. The project contributes to the development of safer agricultural machinery by investigating new sensing approaches for robust perception in harsh outdoor conditions.