UAV Image Clustering Detection of Floating Objects on Floods using Hybrid Firefly Algorithm and Particle Swarm Optimization

Distinguishing floating objects on the surface of the water may be an essential task when placed in the context of disaster responses for floods. This is usually employed by utilizing UAV s that are versatile. Cameras may be equipped to capture images for further assessment. In processing these, nature-inspired approaches have emerged such as the FA and PSO. Utilizing each on its own poses advantages and disadvantages but may be addressed by the proposed hybrid FA-PSO. Specifically, FA is simple and robust but falls short, given it lacks memory in storing the best solution. This is where PSO comes, where it can record both the local and the global best solution and provide faster convergence. The algorithm was tested using images from an aerial dataset for floating objects. Results show that the proposed algorithm was able to obtain faster convergence at the global optimum when compared to its traditional counterpart.