Research on a detection method for blast rock fragments based on concavity-based pre-segmentation.

Accurately detecting the dimensions and particle size distribution of blast rock fragments is crucial for tunnel mucking efficiency and overall construction productivity. To address the challenges of over-segmentation and under-segmentation in existing point cloud segmentation methods for blast rock fragment piles, this study proposed a concavity-based pre-segmentation algorithm. By formulating a concavity criterion, the algorithm improved the spatial separation between adjacent rock fragments. Furthermore, boundary estimation and coplanarity features were incorporated to refine the Euclidean clustering algorithm, resulting in a detection method for blast rock fragment piles driven by concavity-based pre-segmentation. This approach effectively segmented complex blast rock fragment piles, which exhibited significant size variations and irregular surface undulations, into distinct individual fragments. Evidence from practical engineering applications indicates that the proposed method attains an accuracy of approximately 80% in distinguishing the segmentation states of rock fragments. Notably, large rock fragments are effectively segmented without significant cases of over-segmentation or under-segmentation, which verifies the robustness and reliability of the algorithm. In comparison with four state-of-the-art point cloud segmentation algorithms, the proposed approach demonstrates consistently superior performance.