Study On Multi-target Unmanned Rescue Method In Reef Seas

At present,for the rescue of people who have fallen into the reef waters,there are inadequacies such as the inaccessibility of rescue vessels,the inability of traditional lifebuoys to be thrown accurately,the small radius of rescue and the inefficiency of rescue.The current method or equipment for human life saving in reef waters is not perfect enough.This paper designed a multi-targeted unmanned rescue method in reef seas.The shortcomings of traditional lifesaving methods are compensated by the use of drones in conjunction with unmanned lifesavers.In this paper,firstly a rescue scheme of the multi-targeted unmanned rescue in reef waters was designed,and described how it works,and performed related calculations.Secondly,an image recognition algorithm and an environment modeling algorithm were designed for UAVs to capture rescue images,and the algorithm performance tests were conducted.The multi-target path planning algorithm for this rescue method was then designed based on the ant colony optimization(ACO)principle.Finally,simplify the rescue system and complete the experiment of multi-target rescue.Based on the principle of RGB color model,an image recognition algorithm was designed to limit the color range of different objects of the rescue sea images taken by the UAV,so that the position of the relevant objects can be obtained.Environmental map modeling based on location information is then performed for using by path planning algorithms.Based on the ACO principle,a multi-target path planning algorithm is designed for the rescue system,including optimal path calculation,optimal strategy selection,local path optimization and path deviate correction functions.The concept of overall survival rate and survival status of rescued persons is proposed in the optimal strategy selection.The optimal multi-target strategy is selected for rescue by comparing survival rates and then comparing survival status.Experimental work on multi-targeted rescue has been carried out through a simplified rescue system.By analyzing the experimental results,it can be found that this rescue method can realize the functions of image recognition,raster map building,solving the optimal path,selecting the optimal strategy,and local path optimization very well.However,due to the lack of experimental equipment,the walking function could not be completed intact.