Real-time 3D Understanding of Construction Site for Autonomous Robotic Industrial Excavator

In this paper, we propose a set of autonomous excavation systems, which includes the modification of the excavator hydraulic pipeline, the design of a multi-sensor fusion-based perception system to achieve three-dimensional reconstruction of the construction site, the development of a semantic segmentation algorithm for the construction environment to achieve semantic segmentation of the excavation target decision, and the development of a semantic segmentation algorithm for the construction environment to achieve semantic segmentation of the excavation target decision. It is suggested to use a real-time feedback system to make excavating decisions. As a result, the excavation system can perform autonomous perception, decision-making, route planning, and excavation execution tasks.