Efficient monocular pose estimation for complex 3D models

Abstract

Abstract-We propose a robust and efficient method to estimate the pose of a camera with respect to complex 3D textured models of the environment that can potentially contain more than 100, 000 points. To tackle this problem we follow a top down approach where we combine high-level deep network classifiers with low level geometric approaches to come up with a solution that is fast, robust and accurate. Given an input image, we initially use a pre-trained deep network to compute a rough estimation of the camera pose. This initial estimate constrains the number of 3D model points that can be seen from the camera viewpoint. We then establish 3D-to-2D correspondences between these potentially visible points of the model and the 2D detected image features. Accurate pose estimation is finally obtained from the 2D-to-3D correspondences using a novel PnP algorithm that rejects outliers without the need to use a RANSAC strategy, and which is between 10 and 100 times faster than other methods that use it. Two real experiments dealing with very large and complex 3D models demonstrate the effectiveness of the approach.

Backtracking regression forests for accurate camera relocalization

“…It has been widely studied in the context of large scale global localization [10], [11], [6], recovery from tracking failure [12], [13], loop closure detection in visual SLAM [14], global localization in mobile robotics [15], [16], and sports camera calibration [17], [5], [18]. Methods based on local features, keyframes, random forests and deep learning are four general categories of camera relocalization, although other successful variants and hybrid methods [19] exist.…”

Sports Camera Calibration via Synthetic Data

“…Camera Relocalization and Sports Camera Calibration: Camera relocalization has been widely studied in the context of global localization for robots using edge images [6], random forests [7], [8] and deep networks [9], [10], [11], [12].…”

CATCHA: Real-Time Camera Tracking Method for Augmented Reality Applications in Cultural Heritage Interiors

“…As mentioned earlier, three main techniques are available in capturing point clouds in cultural heritage applications, but many model-based camera tracking algorithms require the model to be built using one particular method. For example, in the last decade, an enormous progress in SfM-based location recognition was achieved [8,9]. In addition, some approaches are available that estimate the camera pose based on laser scans.…”