Iterative K-Closest Point Algorithms for Colored Point Cloud Registration

We present two algorithms for aligning two colored point clouds. The two algorithms are designed to minimize a probabilistic cost based on the color-supported soft matching of points in a point cloud to their K-closest points in the other point cloud. The first algorithm, like prior iterative closest point algorithms, refines the pose parameters to minimize the cost. Assuming that the point clouds are obtained from RGB-depth images, our second algorithm regards the measured depth values as variables and minimizes the cost to obtain refined depth values. Experiments with our synthetic dataset show that our pose refinement algorithm gives better results compared to the existing algorithms. Our depth refinement algorithm is shown to achieve more accurate alignments from the outputs of the pose refinement step. Our algorithms are applied to a real-world dataset, providing accurate and visually improved results.

Real-Time Traffic Flow Pattern Matching to Improve Predictive Performance of Online Simulation-Based Dynamic Traffic Assignment

In real-time simulation-based dynamic traffic assignment, selection of the most suitable demand from the library of demands calibrated offline improves the accuracy of the prediction. In the era of data explosion, relying on contextual and rule-based pattern matching logic does not seem sufficient. A rolling horizon scheme for real-time pattern matching is introduced using two pattern matching frameworks. The hard matching algorithm chooses the closest pattern at each evaluation interval, while soft matching calculates the probability of being a match for each pattern. To make sure the pattern switch does not happen because of short-lived interruptions in traffic conditions, a persistency index is introduced. The results show that the number of switches in hard matching is bigger than soft matching but the error of real-time matching for both cases is low. The importance of the results is twofold: First, any observation that is not similar to only one pattern in the library can be mimicked using multiple available patterns; second, more advanced algorithms can match the patterns existing in the library, without any contextual logics for pattern matching.

Relaxation Labeling for Contour Point Matching under Affine Transformations

A matching between two sets of points under affine transformations has attracted more and more attention. Many algorithms devote to extracting the descriptor for the point from the configuration, and the descriptor based point matching is achieved ignoring the pairwise geometric relations. In this paper, taking advantage of the inlier correspondences in matched configurations, we formalize a soft matching criterion which emerges from a matching probability matrix, followed by a relaxation labeling process to refine the match. The proposed approach has been implemented and gives encouraging results under rotation, scaling, shearing and noise.