Research On Vision System Of Fire Fighting Robot Based On Fused Image

Fire fighting robots can replace firefighters in high-risk environments to carry out emergency rescue tasks such as fire extinguishing,and have the potential to be widely used.The accurate and rapid recognition and localization of flames by firefighting robots is the key to improving rescue efficiency,so it is important to improve the flame recognition and positioning effect of the vision system of firefighting robots.This thesis uses a tracked firefighting robot as an experimental platform to investigate the design of a fused vision system for firefighting robots,flame recognition based on fused images,flame localisation based on binocular vision and functional verification of the fused vision system for firefighting robots,with the main research content shown below:(1)Design of integrated vision system for fire robot.Firstly,the fire fighting function requirements of fire fighting robot are analyzed,and the vision system scheme of fire fighting robot based on image fusion is designed.Then,the integrated vision hardware system of fire robot is designed,and related hardware selection is carried out.Finally,the fire robot integrated vision system test bed was built to complete the communication scheme design of the system and provide the basis for the subsequent research.(2)Design of fusion vision system for fire fighting robots.Firstly,the functional requirements of fire fighting robots are analysed,and the scheme of fire fighting robot vision system based on image fusion is designed.Then,the fire fighting robot fusion vision system is designed,and the relevant hardware selection is carried out.Finally,the experimental bench of fire fighting robot fusion vision system is built,and the communication scheme of the system is completed to provide the basis for the subsequent research.(3)Research on flame localization based on binocular vision.Firstly,according to the fire robot flame positioning requirements,the camera calibration theory was analyzed,and a fast calibration scheme was designed based on Zhang’s calibration algorithm.Secondly,a binocular fast calibration method based on manipulator control was proposed.The kinematic modeling and analysis of the 6-DOF manipulator were carried out,and the fast calibration was realized by manipulator control.Then,based on SGBM binocular stereo matching algorithm,parallax map was generated by left and right visible images,and the location of fire area was realized.Finally,the binocular positioning accuracy was analyzed,and the reliability of fast calibration was verified through the comparison experiment between fast calibration and manual calibration.The positioning accuracy of fire robot fusion vision system can meet the fire extinguishing requirements of fire fighting robot.(4)Functional verification of fire fighting robot.integrated vision system.Firstly,based on the functional requirements of the fusion vision system,the binocular fast calibration software and the remote terminal software of the fusion vision system are designed.Then,the field experiment of functional verification of integrated visual system is designed.Finally,through the experimental verification,the system realizes the accurate identification and positioning of fire area,meets the fire fighting operation requirements of fire robot,and has great significance for improving the fire fighting efficiency of fire fighting robot and ensuring the safety of personnel.The thesis concludes with a summary of the work done and an outlook on future research directions for fusion vision systems for fire fighting robot.This thesis contains 85 pictures,28 tables and 89 references.