Stripe Extraction For Structured Light Images And The Applications In Omni-directional Vision System

Structured light vision is widely applied in industrial measurement for its good antijamming,highspeed,accuracy,real-time performance and avoid the feature rectification problem of multi-view image.The study of feature extraction of structured light vision image has import theoretical significance and application value for imporve the accuracy and reliability of visual measurement.By studying the principle of laser structure light visual detection,a structured light visual system were builded,mapping relationship that the position of the target plane and the structure light stripe mapping in the imaging plane were illustrated,the deviation rule of the real center of light stripe in the image were given.This paper also studied the feature extraction of structure light stripe image in the omni-directional structured light vision system,realized the track and predict of the light stripe.The main work of this paper is as follows.The deviation rule were analyzed that the light stripe real center in the image and the center of stripe image with the given measured object,camera and relative posture of the laser plane in the laser structure light vision system.This paper based on the laser structure light vision system model,derived the straight line equation of the edge and center line of the light stripe in the image plane,and gave the modification method of the real center in stripe image on the basis of extract the geometric center of lighe stripe in the image space.The type of catadioptric omni-directional vision system based on spherical mirror and laser emitters were studied,the depth information of the environment by omni-directional structure light vision method were obtained,Using kalman filter to track and forecast the surroundings position,and then reflect it to the image space provide the location reference for feature extraction of the light stripe image.