Digital Twins (DTs) are revolutionizing industrial automation, particularly in intralogistics, by enabling real-time monitoring, predictive maintenance, and process optimization. This paper presents a 3D vision-based DT framework for an Autonomous Mobile Robot (AMR) designed for object localization and management in warehouse environments. The proposed framework integrates a high-fidelity simulation of 3D Time-of-Flight (ToF) cameras to model systematic effects such as multipath interference, flying pixels, and missing pixels. A virtual warehouse, built within the Unity engine, serves as a testbed for evaluating and optimizing object localization algorithms prior to deployment in real-world scenarios. As a case study, we introduce SmartFinder, a vision-based solution for precise pallet detection and localization in automated material handling. By calibrating the virtual ToF camera with real sensor data, we demonstrate the effectiveness of our simulation in predicting measurement uncertainties and improving algorithm robustness. The experimental results validate the DT’s capability to enhance AMR performance, ensuring accurate object localization with an error margin below ± 5 mm. Our findings highlight the advantages of DT-based simulation in reducing testing costs, minimizing operational risks, and accelerating industrial automation deployment. The proposed approach provides a scalable and efficient solution for AMRs operating in dynamic warehouse environments.