Please use this identifier to cite or link to this item:
|Title:||Image-Based Passive Acquisition of Range Data|
|Keywords:||image-based range data;range data;passive acquisition;passive acquisition of range data|
|Abstract:||An image-based technique for passive acquisition of three-dimensional (3-D) range data is proposed. The distance is extracted, in this technique, from the estimation of focus conditions on images produced through a monocular imaging system under natural illumination. The image taken from a 3-D object is generally out-of-focus (defocused). For each surface point, the severity of defocus on the image depends upon how far away the point is from the imaging system and how camera (optical) parameters are adjusted. Each setting of the parameters can be recorded physically, and associated in object-space with the inverse of a distance that corresponds to the position for the sharpest imaging under this setting. Therefore, for a given surface point the defocus severity is a function of such an inverse object-distance. It can be shown that this function is symmetrical to, and monotonic on both sides of, a point corresponding to the inverse distance of the surface point. To estimate the parameters of the function (one of which is the inverse distance of the surface point), 3~4 images need to be taken under different camera settings with known associated inverse distances in object-space, determined through a once-for-all calibration procedure. Defocus severity is evaluated from a calculation on the window image that corresponds to a small area around the surface point, and the inverse variance in the window is suggested in this technique for the best performance. The 3-D surface geometry is acquired by applying the algorithm, in parallel, to all surface points in the field of view. Various aspects of the technique are discussed and several algorithms are developed. The technique is implemented on an opto-digital imaging system and evaluated under different conditions. A number of objects are tested to demonstrate its performance.|
|Appears in Collections:||Digitized Open Access Dissertations and Theses|
Files in This Item:
|xu_shi_1992_masters.pdf||4.79 MB||Adobe PDF||View/Open|
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.