Font Size: a A A

The Research Of Displaying And Mosaicing Technology For Sonar Image

Posted on:2013-10-31Degree:MasterType:Thesis
Country:ChinaCandidate:J GuanFull Text:PDF
GTID:2248330377458583Subject:Underwater Acoustics
Abstract/Summary:PDF Full Text Request
This paper has mainly designed a displaying software for sonar. The function of thesoftware is to display panoramic-joining target. It is suitable for the situation that the target istoo big for sonar to take one-time image.This software uses network port to communicate with the underlying hardware for datatransmission, and the data will be real-time displayed.This paper has focused on image matching algorithms. The main research is about SIFTfeature points’ detection and registration technology, and Harris corner points’ detection andNCC registration technology. To compare the two algorithms in advantages and disadvantages,this paper will discuss their respective applicability.For SIFT algorithm, here are two improvements. One is to use the affine transformalgorithm to improve the original registration algorithm and make it better in resisting affinetransformation in order to solve the influence of sonar’s working environment. Anotherimprovement is to improve the strategy of detecting feature points so that it can do better indetecting feature points.For NCC registration algorithm here is also an improvement, which is to change theoriginal fixed threshold into adaptive threshold. It mainly uses the two images’ affinetransformation to search this two images’ max mutual-information entropy, and then find theoptimal threshold. Through the improvement this algorithm can do better in getting matchingfeature points, and will avoid the mistake made by original fixed threshold which is pairingless or mistaken points, etc.This paper has also studied other algorithms associated to image stitching, such as thecylindrical projection algorithm、the image smoothing technology and so on.
Keywords/Search Tags:sonar image, image registration, image stitching, feature points’ detection
PDF Full Text Request
Related items