Font Size: a A A

Research On Key Technology Of Stereoscopic Microsurgery System Based On Binocular Vision

Posted on:2019-08-15Degree:MasterType:Thesis
Country:ChinaCandidate:X S LiuFull Text:PDF
GTID:2404330596450167Subject:Aviation Aerospace Manufacturing Engineering
Abstract/Summary:PDF Full Text Request
In recent years,the development of stereovision technology is remarkable,and binocular stereo vision is the most widely used.Binocular stereo vision through two camerasfrom different angles capture the same scene,and got two images with disparity,thenuse the disparity informationto restore the three-dimensional information.It plays an important role in fields as industrial detection,automobile navigation,non-contact distance measurement,virtual vision and so on.The application of binocular vision is often seen in the medical field,such as surgical navigation,internal contrast,and so on.Combine binocular vision and 3D display,and collectionthe images from operation microscope,thenshow them to doctor by high definition 3D display after some processing,this technology called stereoscopic microsurgery.The application of binocular vision technique to microsurgery is the same as the traditional binocular vision principle,but the accuracy and speed havehigher requirements.Based on the above content,this paper elaborated the basic principles of binocular vision,from the principle of camera imaging to the triangulation principle,and to the stereo matching technology,especially the stereo matching technology.Based on the original adaptive support-weight matching algorithm,by changing the shortcoming of assigning weights only based on the characteristics of the image itself.In order to further improve the matching precision of the occlusion region,the steps of the right and left consistency check and median filter are added to the algorithm.Then,for the speed of the algorithm,using the parallel computing power of the GPU platformto parallelizing the cost calculation and cost aggregation.Testing the standard image pairs from the Middlebury platform.The results show that the improved algorithm is more than 10% higher than the original algorithm in accuracy,and the speed of algorithm is also greatly improved based on GPU and optimization,has the applicability of the stereoscopic microsurgical systemThe binocular vision system is built,and the software is written based on OpenCV and VS2010.The whole process of image acquisition to 3D display is realized.The experimental results showsthe calibration results are highly accurate,and 3D displays has good stereoscopic effect.Finally,we use the improved algorithm to extract the depth information.Compared with the actual depth shows that most regions have relatively small depth information errors,which verify the reliability of the system.
Keywords/Search Tags:binocular vision, stereo matching, adaptive weight, GPU, parallel computing
PDF Full Text Request
Related items