| Rescue detection robot is a mine emergency rescue equipment.After coal mine accident,rescue detection robot can carry a variety of sensors into the disaster area to complete the dangerous and complex rescue mission.In this paper,the machine vision technology is used to realize the visual perception function of the robot.The distribution information of the obstacles in the robot field is obtained by image processing and analysis,and the robot path planning problem under the complex environment structure is studied.Specific work has been done as follows:Firstly,the paper describes the background and significance of the research on the research of the underground rescue robot.It introduces and analyses the present situation and development trend of the machine vision technology and path planning.It also analyses the classification and application field of the existing path planning method.Secondly,the paper introduces the principle of digital image stabilization technique,simulates the underground image of the underground roadway,whose stable images are processed by the image preprocessing,image segmentation,image morphological processing and obstacle extraction.And it calibrates the camera to derive the distance formula of the robot and the obstacles.Thirdly,it emphatically analyses the process of robot path planning and its implementation,including Dijkstra algorithm for rapid generation of sub-optimal collision avoidance path.Two new improved intelligent algorithms are introduced,namely particle swarm algorithm and ant colony algorithm.Fourthly,it analyses and evaluates the advantages and disadvantages of PSO algorithm and ACO algorithm.The fusion process of two algorithms is introduced,and an improved fusion algorithm is proposed.By establishing MATLAB simulation environment,PSO algorithm,improved ACO algorithm and improved fusion algorithm to test the simulation analysis,analysis of the obtained parameter data,then it draws conclusions. |
PDF Full Text Request