Research on Simulation Experiment of UAV Crossing Fire Circle Based on AirSim

This article uses AirSim for simulation and modeling to address the use of unmanned aerial vehicles (UAVs) in the field of firefighting. The firefighting process is simplified into four different modules: “water collection,” “target identification,” “target traversal,” and “return and landing.” In the process of implementing “target identification” and “target traversal,” the article utilizes AirSim's camera module and the OpenCV library. It employs ROI region binarization and BP techniques to identify numbers and uses the Hough transform technique to detect the coordinates of the fire circle and the relative position of the UAV to the fire circle. Ultimately, the UAV successfully completes the aforementioned four tasks, providing a feasible theoretical algorithm for UAVs to perform more complex real firefighting missions. The article concludes with a preliminary evaluation of the performance of this algorithm and proposes areas of improvement and future work.