Research On Monocular Vision Of NAO Robot And Its Application In Object Search And Rescue

With the continuous development of science and technology,robots gradually into people's sight,as one of the most widely used robots in the world,the monocular vision of NAO has become one of the most popular research fields.Taking the NAO robot as the development platform,based on machine vision,building a maze,simulating object search and rescue scene,implementation in the labyrinth of unknown environment,finding objects,grabbing objects,and selecting the best return path,bring the object out of the maze.This paper first introduces the NAO robot hardware,software environment,visual system and its research status,then the path planning,deviation correction,target recognition and crawling,and optimal return path are studied.The main work of this paper is divided into the following four parts:In terms of path planning,because the location of the object in the maze is unknown,the two kinds of path planning algorithm,therefore,by analyzing the two commonly used path planning algorithm,namely the depth-first search algorithm and breadth-first search algorithm,the depth first search algorithm is selected to search the object.In terms of deviation correction,NAO robot in the maze due to the motor itself error,fever and other factors will cause the deviation of the walking,Eventually causing the robot to hit the maze wall,which will affect the normal walking of the robot,therefore,a correction method of course correction and phase shift of the body is proposed.In terms of target recognition and crawling,through the threshold to get the binary image of the object,and calculate the contour of the perimeter,area.Setting the threshold to determine the perimeter,area and perimeter of the square and the area ratio,when the three meet the threshold conditions,the contour is the contour of the target object.Calculation of the center of the object position in the image,the position and angle adjusting robot,and judge object center in the location of the image,so that the object is located at a specific position in the view of camera.When the center of the object is within the graspable range in the image,a pre-set grasping action is carried out to effect grasping of the object.In terms of the optimal return path,when the path planning is carried out,the maze information is stored continuously,Therefore,when the maze of new path information,it is added to the list,when there is a dead corner,the corresponding adjustment,and accordingly to delete and add on the list.According to the law,when the object is finally found,the shortest path is stored in the list,after the completion of the object,by tracing operation,constantly extract the list in the last list,determine the direction,and the corresponding movement,that is,the use of the shortest path to find the object out of the maze.In this paper,Python is used to write the program,the experiment is carried out by NAO robot,implementation to find object,fetching in the maze,select the optimal path to return.the effectiveness and robustness of the algorithm are verified by analyzing the experimental results.