Research On Obstacle Avoidance Navigation System Of A Mobile Platform Based On Stereo Vision

In some situations where there are obstacles,dangerous goods such as ammunition will be stored.It is more dangerous to carry these goods by manpower.To choose an independent transport vehicle to operate these dangerous goods can greatly ensure the safety of personnel.In order to realize the autonomous operation of transport vehicle in this kind of environment,this article is based on stereo vision technology,researching on autonomous obstacle avoidance navigation system of a mobile platform,providing a convenient and safe way for the transportation of dangerous goods in these special scenes.Binocular stereo vision technology has the advantages of low power consumption,low cost,simple use,easy maintenance and upgrade.The research based on this will help to improve the transport efficiency of dangerous goods and ensure the safety of personnel to the greatest extent.In this paper,how to use binocular stereo vision technology to realize the autonomous operation of transport vehicle in special situations is studied.1.In order to meet the security requirements of the system in special scenes,it is necessary to get accurate binocular camera imaging model.The purpose,principle and method of camera calibration are described.The binocular camera is calibrated and the results are given.According to the calibration results,the accurate imaging model of the binocular camera is obtained.Secondly,it introduces three image enhancement methods used in this paper,which are also helpful to the security of the system.Finally,the stereo matching algorithm in stereo vision system is introduced.Two typical stereo matching algorithms are analyzed and a contrast experiment is designed to compare their ability in matching effect,matching speed and other aspects,and the matching algorithm is selected with the consideration of system security and real-time.2.According to the actual working environment of the transport vehicle,the characteristics of the depth information collected and the actual needs of obstacle avoidance navigation,this paper designs a series of depth information post-processing work,which transforms the depth information into intuitive information that can be used by the system obstacle avoidance unit.A three-dimensional reconstruction method based on coordinate transformation is proposed,which can reconstruct the collected depth information in three dimensions.For single or multiple obstacle scenes,the collected depth information is fused in three dimensions to generate three-dimensional map,which is used as the navigation information of subsequent transport vehicles to realize the path planning function.On the basis of analyzing the characteristics of binocular stereo vision work,this paper designs an ARM board as the binocular stereo vision module control core system scheme.Briefly introduce the core board parameters and main image processing programs.The library functions and programs were cross-compiled,transplanted,and debugged.The system was finally run on the mobile end.3.The principle of artificial potential field method is briefly described.Aiming at its own defects,combined with the obstacle avoidance information,i.e.the processed depth map,a variety of improvement measures are proposed.The dynamic window method is also introduced,which takes the three-dimensional map as the navigation information and expands the types of obstacles in the algorithm according to the actual work scene to complete the navigation work.4.Based on the analysis of the characteristics of binocular stereo vision technology and working environment,the paper designs a system scheme with arm board as the control core of binocular stereo vision module.The core board parameters and main image processing programs are briefly introduced.The library functions and programs are cross compiled,transplanted and debugged to realize the system running on the mobile terminal.5.According to the research requirements,this paper designs and completes the experiment of distance measurement and positioning of binocular stereo vision unit to test the security of the system;builds a test prototype to carry out comprehensive obstacle avoidance navigation experiment,and verifies the key technologies of the system.In this paper,a mobile platform obstacle avoidance navigation system designed in a special environment can collect real-time and accurate environmental depth information,achieve obstacle avoidance work,effectively and quickly achieve the three-dimensional reconstruction of environmental information,according to the reconstructed three-dimensional obstacle map for navigation.The system has the advantages of safety,reliability,low cost and convenient use.It can not only be used for the transportation of dangerous goods in the scene of obstacles,but also has certain reference significance for the transportation vehicles in industrial production.