Towards Autonomous Mobile Crushing: Single-particle Detection, Real-time Process Monitoring, and Decision Support

Abstract This contribution presents the development and experimental validation of a laboratory-scale Digital Twin Camera System for image-based particle size distribution (PSD) estimation in mobile impact crushing. The system reproduces the crusher discharge zone under controlled laboratory conditions and enables a systematic comparison between camera-derived PSD results and reference sieve analyses. The setup integrates synchronised industrial cameras, controlled illumination, calibration routines, and a web-based monitoring interface to generate reproducible datasets for optical analysis. Experimental investigations with five representative material types showed good agreement between camera-based and sieved PSD in the coarse and intermediate fractions, between 8 and 30 mm, while larger deviations occurred in the fine fraction below approximately 8 mm due to optical detection limits. Additional tests with red optical filters improved the contrast and segmentation stability, particularly for materials with complex surface structures such as asphalt. The Digital Twin environment also enabled the identification and elimination of data transfer and connectivity issues that could not be resolved during field operation. As a result, a stable and reliable camera monitoring workflow for determining PSD on a conveyor belt was established, preparing the system for future industrial deployment in mobile crushing operations.