Research On Image Mosaic Technology In Robot Vision

Video mosaic is an important research field based on image mosaic,which can be used in social security,medical image,self-driving,tourism and other fields.Aiming at the limitation of monocular vision of mobile robot,this paper proposes multi-camera video splicing to replace monocular vision with panoramic vision.This paper first describes the relationship between video mosaic and image mosaic,from image mosaic deduction to video mosaic.In the real-time problem of image registration,the advantages and disadvantages of SIFT and SURF algorithms are analyzed.On the basis of this,an improved SURF algorithm with stronger real-time performance is proposed,that is to say,when extracting feature points,piecewise linear processing is applied to the image.The inconspicuous feature points are suppressed,and the detection region is halved,and the generated feature vectors are reduced by principal component analysis.The simulation results show that the improved SURF algorithm not only improves the real-time performance,but also preserves it.The scale invariance and rotation invariance of the original algorithm are good.Secondly,the details of the fusion region will be blurred because of the improper selection of the weights in the detail part of the fusion region,so this paper improves the method based on the gradual in-out fusion method,which is based on the step-in-step-out fusion method.The feature points and the significant pixels in the neighborhood are preserved,and the non-characteristic points are fused by the conventional method of gradual in and out.This method can also highlight the details of splicing on the basis of eliminating stitching seam.Finally,the algorithm of image mosaic is deduced in video mosaic,and the key frame is selected in video mosaic.The selection of keyframes is based on their adjacent two The pixel difference of the frame image is determined and the number of frames selected per second is no less than the minimum number of frames that the human eye can see.In order to test its real-time performance in a specific environment,FPGA is used to capture the video,and the collected video is transmitted to DSP.The algorithm is implemented in DSP.In summary,the improved algorithm proposed in this paper has a higher efficiency in video mosaic,and the details of the splicing image are more prominent,which can be applied in practice.