Worldwide, streams and lakes are important resources for human survival and development (Giuliani et al., 2022;Sauer et al., 2021;Vörösmarty et al., 2010). Topological systems can reflect the connectivity and flow rules between streams and lakes, which is the basis for simulating the process of material and energy transport in stream network systems (including streams and lakes) (Dai et al., 2021;Verdin, & Verdin, 1999). In some multilake areas, especially inland plateaus, lakes play an important role in stream network systems; therefore, constructing the topological systems between streams and lakes is of great interest to hydrologists, climate scientists and geomorphologists.Flow direction is the key factor in constructing a lake-stream topological system, as flow direction provides natural routing information for streams and lakes, allowing for the organisation of stream networks based on certain physical rules. Accurate flow direction is the basis for many geoscience studies, such as hydrological modeling, Earth's surface process simulation, and terrain classification (C. Z. Qin et al., 2011). Although geoscience studies relating to stream networks have greatly contributed to promoting the development of geography, the basic flow direction calculation is still not realistic in many situations (Gutierrez & Abad, 2014;Whipple & Viers, 2019). This problem is due to the lack of objects during the process of flow direction calculation. However, the Earth's surface is a complex system (S. Li et al., 2020;. There is a significant difference in the flow process of stream networks under different underlying surface conditions. Regardless of the complexity of relevant geoscience studies, object-oriented flow direction algorithms can improve the reliability of flow direction calculations.The D-infinity algorithm, as a widely used traditional flow direction algorithm, uses the maximum downslope of a 3 × 3 window as the flow direction value; however, this algorithm performed poorly in flat areas (Tarboton, 1997).