Simultaneous localization and mapping (SLAM) is an emerging robotic technology enabling autonomous robots to self-localize and map the surrounding environments. However, SLAM system is time and resource consuming, making it hard to implement on mobile devices. In this work, ROPY-SLAM, a novel heterogeneous framework for SLAM is proposed to take advantage of various devices such as personal computer and FPGA. The proposed system reduces the execution time by 28.6% and resource utilization by at most 71.2%. Moreover, ROPY-SLAM supports more flexible hardware and software implementations, providing a solution for distributed calculation of robotic applications.