Navigation Technology Snake-arm Robot Vision Positioning

Vision based localization and navigation technology is an important researchdirection in the field of intelligent robots, which builds a description of the robotssurroundings. Vision based localization and navigation also known as visual servo, isa technology which uses feedback information extracted from a vision sensor tocontrol the motion of a robot. As snake-arm robot’s working environment is a properunstructured and unknown three-dimensional space, a lot of vision based localizationand navigation methods applying in the ground mobile robot platform obtain littlegood application in the snake-arm robot platform, which makes the study of visionbased localization and navigation technology more challenging.This paper focus on the study of vision based localization and navigationtechnology based on the localization and navigation mission requirements forsnake-arm robot in the confined space in some aviation industry’s application. Themain contents of this paper are as follows:(1) A review of the binocular stereovision system’ key technology in detail ispresented, include camera calibration, image distortion correction, stereo rectificationand three-dimensional reconstruction, and related experiments are carried out in thispaper.(2) An active binocular stereo vision obstacle detecting system is proposed,which combined stereo vision system with a line structured light. Related obstacledetecting experiments are designed to confirm the obstacle detecting system’effectiveness and real-time feature, especially for the obstacles with poor texture.(3) The methods research of target recognition and localization based on imagesegmentation is carried out on the Mean-Shift image segmentation. For targetrecognition: first, segment images and mask segmentation regions, then extract imagefeature information and recognize targets by feature matching at last. For targetlocalization: extract and track the edge of the object image region, then build thedisparity of object contour and acquire real-time location information about the targetusing binocular stereo vision measurement system.