Font Size: a A A

Research On 3D Surface Shape Recovery Of Objects Based On A Single Image

Posted on:2006-10-02Degree:MasterType:Thesis
Country:ChinaCandidate:G J FanFull Text:PDF
GTID:2168360155959947Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
The research on 3D surface shape recovery of objects is an important field in computer vision. At present, many recovery techniques have been developed, in which the method named shape from shading (SFS) makes use of the change of image brightness to recover such parameters as the relative height and surface normal. It is one of the key techniques of 3D shape recovery in computer vision field. This method is noticeable because it is not only easily realized but also applicable for the condition in which other methods are not suitable.The method of 3D shape recovery mentioned in this paper can gain 3D dimensions of objects in 0~180° only using a single image on the basis of brightness change. In this paper, the background knowledge and the methods related to SFS are briefly introduced. Based on the analysis and comparison of various reconstruction algorithms, this paper gives a new method of 3D surface shape recovery of objects from a single image. It is more simple and applicable. Using this method, hardware equipments only include image pick-up system and a computer. The operation mode is the uncomplicated principle of optical lens imaging. The core idea of this research is using image shading to recover 3D dimensions of objects. The shading is the 3D cues left in 2D image. In the coordinates of light source, the surface normal vectors corresponding to every pixel are calculated. Then these vectors are converted to the coordinates of image through the coordinate conversion, so 3D dimensions of objects are gained. This paper uses several experiment objects to validate the feasibility of this method and the results of 3D recovery are simulated in MATLAB. This paper also analyzed the precision of the 3D recovery results.The method introduced in this paper can be used in different fields. It has a favorable application prospect in iatrology, weather, remote sensing, especially in national defense and military affairs. It also can be used for an assistant analysis method of other 3D reconstruction methods and help analyze...
Keywords/Search Tags:SFS, Lambertian illumination model, 3D recovery, image processing, surface normal, precision analysis
PDF Full Text Request
Related items