Variable baseline/resolution stereo

Abstract

We present a novel multi-baseline, multi-resolution stereo method, which varies the baseline and resolution proportionally to depth to obtain a reconstruction in which the depth error is constant. This is in contrast to traditional stereo, in which the error grows quadratically with depth, which means that the accuracy in the near range far exceeds that of the far range. This accuracy in the near range is unnecessarily high and comes at significant computational cost. It is, however, non-trivial to reduce this without also reducing the accuracy in the far range. Many datasets, such as video captured from a moving camera, allow the baseline to be selected with significant flexibility. By selecting an appropriate baseline and resolution (realized using an image pyramid), our algorithm computes a depthmap which has these properties: 1) the depth accuracy is constant over the reconstructed volume, 2) the computational effort is spread evenly over the volume, 3) the angle of triangulation is held constant w.r.t. depth. Our approach achieves a given target accuracy with minimal computational effort, and is orders of magnitude faster than traditional stereo.

Efficient and Globally Optimal Multi View Dense Matching for Aerial Images

“…This means that depth precision is mainly a function of the ray intersection angle (Gallup et al, 2008) α, having a minimum at α = 90…”

Consistent Depth Maps Recovery from a Video Sequence

“…In order to improve the final reconstruction quality, they used optical flow to find corresponding pixels in the subsequent frames of the same camera, and enforced the temporal consistency in reconstructing successive frames. With the observation that the depth error in conventional stereo methods grows quadratically with depth, Gallup et al [14] proposed a multibaseline and multiresolution stereo method to achieve constant depth accuracy by varying the baseline and resolution proportionally to depth.…”

An Integer Linear Programming Model for View Selection on Overlapping Camera Clusters

“…Tingdahl et al [26] reduce the set of initial views relying on depth maps data. Goesele et al [9] and Gallup et al [8] select viewpoints relying on simple properties of the input images such as resolution or baseline. Ladikos et al [13] propose a spectral clustering approach which incorporates scene and camera geometry to build a similarity matrix and then use mean shift to automatically select the number of clusters.…”

“…The linear constraints enforce the shared visibility of points between cameras to guarantee the coverage of the final 3D reconstructions. Solving the ILP model we find a globally optimal set of cameras, avoiding any heuristics or greedy iterative procedures used in other works [26,9,8,6].…”