C. Matsunaga and K Kanatani. Calibration of a moving camera using a planar pattern: optimal computation, reliability evaluation, and stabilization by model selection. In Proc. 6th European Conference on ComputerVision(ECCV'00), Dublin, Ireland, July 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....with these algorithms is that they minimize algebraic distances and a good choice of system normalization (as we will discussed later) is often necessary to obtain reliable results. When only the focal length can vary, an optimal solution (attaining the Cramer Rao lower bound) is proposed in [5]. Nevertheless, it requires an off line pre calibration step (not described) and Newton iterations in order to only consider the different unknown focal lengths uncalibrated. Motivation. Within the framework of plane based calibration of a camera with varying internal parameters, our aim is to ....

C. Matsunaga and K Kanatani. Calibration of a moving camera using a planar pattern: optimal computation, reliability evaluation, and stabilization by model selection. In Proc. 6th European Conference on ComputerVision(ECCV'00), Dublin, Ireland, July 2000.

....Application Fig. 6(a) is a real image of a toy, behind which we placed our optimal grid pattern. After segmenting the toy image from the background by using a chromakey technique, we computed the 3 D position and focal length of the camera by observing an unoccluded portion of the grid pattern [2, 4]. The focal length is estimated to be 576 pixels. The standard deviations of the focal length, the translation, and the rotation are evaluated to be 4 38.3 pixels, k5.73cm, and :k0.812 , respectively. Fig. 6(b) is the top view of the estimated camera position and its uncertainty ellipsoid (three ....

C. Matsunaga and K. Kanatani, Calibration of a moving camera using a planar pattern: Optimal computation, reliability evaluation and stabilization by model selection, Proc. 6th Euro. Conf. Computer Vision, June 2000, Dublin, Ireland, Va. 2, pp. 595- 609.

....very slowly. These problems have been dealt with by ad hoc measures in the past. We now show that they can be resolved by model selection: we model various modes of camera motion and zooming that ax e likely to occur in practice and choose at each frame the most appropriate one by model selection [11]. Fig. 5 shows five sampled frames from a real image sequence. Here, we used the following models (see [11] for the details of the computation) The camera is stationary with fixed zooming. The camera rotates with fixed zooming. The camera lineax ly moves with fixed zooming. The camera ....

....be resolved by model selection: we model various modes of camera motion and zooming that ax e likely to occur in practice and choose at each frame the most appropriate one by model selection [11] Fig. 5 shows five sampled frames from a real image sequence. Here, we used the following models (see [11] for the details of the computation) The camera is stationary with fixed zooming. The camera rotates with fixed zooming. The camera lineax ly moves with fixed zooming. The camera moves arbitrarily with fixed zooming. The camera moves arbitrarily with linearly changing zooming. ....

C. Matsunaga and K. Kanatani, Calibration of a moving camera using a planar pattern: Optimal computation, reliability evaluation and stabilization by model selection, Proc. 6th Euro. Conf. Cornput. Vision, JunesJuly 2000, Dublin, Ireland, to appear.

....plays an important role as well as its degree of freedom [4, 5] Lines and curves in two dimensions all have dimension 1. Here, we compare two models that have different dimensions: we fit a line and a plane to points in three dimensions. We randomly take eleven points in a rectangulax re gion [0, 10] x [ 1, 1] in the xy plmm and enlarge them A times in the y direction. Adding rmdom Gaussian noise of stmdard deviation rr to the x, y, md z coordinates of each point independently, we fit a line and a plane in a statistically optimal manner (see [4] for the details of the computation) Note that ....

....object and sepax ate the object image from the pattern image by a chromakey technique. Since the true geometry of the grid pattern is known, the position md the focal length of the camera can be determined if four or more grid points are detected in the unoccluded part of the grid pattern image [10]. However, the following two problems must be resolved: 1. When the camera optical axis is perpendicular to the pattern, the 3 D position and focal length of the camera are indeterminate because zoom ing out and moving the camera forwax d cause the same visual effect. 2. Since some unoccluded ....

C. Matsunaga and K. Kanatani, Calibration of a moving camera using a planar pattern, Proc. 5th Symposium on Sensing via Image Information, June 1999, Yokohama, Japan, pp. 255 260.

No context found.

C. Matsunaga and K Kanatani. Calibration of a moving camera using a planar pattern: optimal computation, reliability evaluation, and stabilization by model selection. In Proc. 6th European Conference on ComputerVision(ECCV'00), Dublin, Ireland, July 2000.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC