Research On Video Mosaic Technology Of Tetrahedral Robots

Compared with traditional mobile robots,tetrahedral robots have better motion stability and environmental adaptability.It shifts the position of the center of gravity through the deformation of its own mechanism,and then the overall instability occurs and rolls,which puts forward higher requirements for its image perception system.Aiming at the problem that the field of view of the tetrahedral robot equipped with the camera is small,the video mosaic technology is researched,and two video mosaic schemes in the initial state and rolling gait are proposed and verified by experiments.Firstly,in view of the serious distortion of the image captured by the fisheye camera on the tetrahedral robot,the principle of the fisheye camera imaging model is analyzed,and the transformation relationship between the three-dimensional space point coordinates in the environment and the two-dimensional pixel coordinates of the imaging plane is obtained.Through the checkerboard calibration board,the camera internal parameter coefficient and fisheye lens distortion coefficient are solved,and on this basis,the corrected image is solved,and an image with a wider field of view and no obvious distortion is obtained.Secondly,in order to increase the range of the image field of view and reduce the blind area field of view,this paper studies the image stitching algorithm,analyzes the principle of feature point extraction of the SURF algorithm and the ORB algorithm,and compares the extraction number and the number of matches after screening.It is concluded that using the SURF algorithm to extract feature points can better ensure the registration accuracy and speed.After image registration,a stitched image with no obvious stitching,stitching dislocation and smooth brightness transition is obtained by weighted average image fusion.Thirdly,this paper introduces a key frame extraction algorithm in the process of video mosaic,and extracts key frames through the inter-frame difference method.The image registration and spatial transformation relationship are only calculated for the image to be stitched at the moment of the key frame,and the spatial transformation relationship calculated at the previous key frame moment is used at other moments,which effectively improves the processing speed of video mosaic and ensures the continuity and real-time of the video.to reduce the judder in the video.Finally,this paper applies the video mosaic technology to the tetrahedral robot for experiments,and proposes two video mosaic schemes.Among them,in the initial state,the video images collected by the horizontal field of view of the camera at the top of the robot are stitched to output a spliced video with a larger field of view,reducing the operation of switching the angle of view and reducing the blind spot of the field of view;in the rolling gait,the direction of movement is adjusted.The video images collected in the vertical field of view of the adjacent cameras are stitched,and the image with a stable field of view is output after cropping,and the thing in front is always in the center of the image.