<p>The unmanned transportation in open-pit mines relies heavily on efficient and safe path planning for mining trucks. To reduce excessive detours of mining trucks in unstructured areas, innovative research is conducted on a three-dimensional path planning method with crossing and bypassing collaboration. Firstly, the effective point cloud data after filtering and registration are gridded, and an environment map containing obstacle height information is obtained based on the digital elevation model. Secondly, a three-dimensional collision detection method and a crossing and bypassing collaborative passage strategy are established, which can effectively balance safety and passage efficiency. Then, a fusion algorithm based on improved ant colony and improved artificial potential field is proposed, which can safely avoid dynamic obstacles while shortening the path length. Finally, an unmanned mining truck and an unstructured scene are set up to implement path planning experiments. The results show that the unmanned mining truck equipped with the fusion algorithm enables selective obstacle crossing, reducing path length by 31.90% and actual driving time by 32.51%, with higher path smoothness. In addition, the mining truck can reasonably plan local paths and accurately avoid dynamic obstacles. This verifies that the proposed three-dimensional path planning method with crossing and bypassing collaboration has good global optimization performance and local dynamic avoidance capability, which can significantly improve operational efficiency and safety.</p>