Shape Recognition of Material Heaps in Outdoor Environments and Optimal Excavation Planning

For the realization of an automated material heap excavation system in a recycling plant, an algorithm for its planning process is proposed. The approach uses depth and color data from a 3D stereo camera. Based on the obtained information, the material heap is recognized and segmented. The resulting segments are evaluated according to a reward function for optimality in terms of efficiency, safety and practicability for excavation. While this approach is developed for recycling plants in particular, other applications can be expected.