Automatic Dense Reconstruction from Uncalibrated Video Sequences. Front Cover. David Nistér. KTH, – pages. Automatic Dense Reconstruction from Uncalibrated Video Sequences by David Nister; 1 edition; First published in aimed at completely automatic Euclidean reconstruction from uncalibrated handheld amateur video system on a number of sequences grabbed directly from a low-end video camera The views are now calibrated and a dense graphical.
|Country:||Antigua & Barbuda|
|Published (Last):||23 June 2007|
|PDF File Size:||3.25 Mb|
|ePub File Size:||13.11 Mb|
|Price:||Free* [*Free Regsitration Required]|
Maxime Lhuillier’s home page
The running times of the algorithm are recorded in Table 2and the precision is 1 s. The first involves the SfM calculation of the images in the queue.
Table 1 lists all of the information for the experimental image data and the parameters used in the algorithm. Small differences in the parameters between the subregions will result in discontinuous structures. The main text gives a detailed coherent account of thetheoretical foundation for the system and its components. First, a principal component analysis method of the feature points is used to select the key images suitable for 3D reconstruction, which ensures that the algorithm improves the calculation speed with almost no loss of accuracy.
When we use bundle adjustment to optimize the parameters, we must keep the control points unchanged or with as little change as possible. Yufu Qu analyzed the weak aspects of existing methods and set up the theoretical framework. Incremental smoothing and mapping using the Bayes tree. For the pot experiment, most distances are less than 1. In order to test the accuracy and speed of the algorithm proposed in this study, real outdoor photographic images taken from a camera fixed on a UAV and standard images together with standard point cloud provided by roboimagedata [ 27 ] are used to reconstruct various dense 3D point clouds.
The patch-based matching method is used to match other sequneces between images. Accurate, dense, and robust multiview stereopsis. To ensure the smooth of two consecutive point cloud, an improved bundle-adjustment named weighted bundle-adjustment is used in this paper. These methods can improve the speed of the structure calculation without loss of accuracy. These contributions include the work of Liu et al. In order to test the speed of the proposed algorithm, we compared the time consumed by our autpmatic with those consumed by openMVG and MicMac.
Then, the singular value decomposition SVD of matrix A yields two principal component vectors. A paradigm for model fitting with applications to image analysis and automated cartography.
After that, a dense point data cloud and mesh data cloud can be obtained. Equation 9 is the reprojection error formula of the weighted bundle adjustment. In addition, the algorithm must repeat the patch expansion and point cloud filtering several times, resulting in a significant increase in the calculation time. We delete k images at the front of the queue, save their structural information, and then place k new images at the tail of the queue; these k images are then recorded as a set C k.
Finally, a dense 3D point cloud can be obtained using the depth—map fusion method. Although unfalibrated and k are fixed and their values are generally much smaller than Nthe speed of the matching is greatly improved.
Rapid 3D Reconstruction for Image Sequence Acquired from UAV Camera
When processing weakly textured images, it is difficult for this method to generate a dense point cloud. The flight distance is around m. This is achieved by weighting the error term of the control points. Application of open-source photogrammetric software MicMac for monitoring surface deformation reconstrucyion laboratory models.
If the two images are captured almost at the same position, the PCPs reconztruction them almost coincide in the same place. Positions and orientations of monocular camera and sparse point map can be obtained from the images by using SLAM algorithm.
When calculating the structure by the queue, optimization of the bundle adjustment causes the parameters to reach the subregion optimum rather than the global optimum. It usually returns a completely wrong estimate. On an independent thread, the depth maps of the images are calculated and saved in the depth-map set.
Then, the structure of the images in the queue is computed, and the queue is updated with new images. The depth maps are optimized and corrected using the pixel matching algorithm based on the patch. The second step involves obtaining the 3D topography of the scene captured by the images.
To compress a large number of feature points into three PCPs Figure 2 b. Journal List Sensors Basel v. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution CC BY license http: Convex relaxation is proposed by some authors to avoid convergence to local minima. First, we use the scale-invariant feature transform SIFT [ 19 ] feature detection algorithm to detect the feature points of each image Figure 2 a.
The UAV flight over the top of the buildings. The corresponding image pixels of P r are marked as a set U rand the projection relationship is expressed as P: Conflicts of Interest The authors declare no conflict of interest.
In order to complete the dense reconstruction of the point cloud and improve the computational speed, the key images which are suitable for the structural calculation must first be selected from a large number of UAV video images captured by a camera. The first two terms of radial and tangential distortion parameters are also obtained and used for image rectification. The first step involves recovering the 3D structure of the scene and the camera motion from the images.
Automatic Dense Reconstruction from Uncalibrated Video Sequences | Open Library
The final results accurately reproduce the appearance of the scenes. The flight height is around 80 m and is kept unchanged. In order to test the accuracy of the 3D point cloud data obtained by the algorithm proposed in this study, we compared the point cloud generated by our algorithm PC with the standard point cloud PC STL which is captured by structured light scans The RMS error of all ground truth poses is within 0. The process steps are as follows.