We present MOPED, a framework for Multiple Object Pose Estimation and Detection that seamlessly integrates single-image and multi-image object recognition and pose estimation in one optimized, robust, and scalable framework. We address two main challenges in computer vision for robotics: robust performance in complex scenes, and low latency for real-time operation.We achieve robust performance with Iterative ClusteringEstimation (ICE), a novel algorithm that iteratively combines feature clustering with robust pose estimation. Feature clustering quickly partitions the scene and produces object hypotheses. The hypotheses are used to further refine the feature clusters, and the two steps iterate until convergence. ICE is easy to parallelize, and easily integrates single-and multi-camera object recognition and pose estimation. We also introduce a novel object hypothesis scoring function based on M-estimator theory, and a novel pose clustering algorithm that robustly handles recognition outliers.We achieve scalability and low latency with an improved feature matching algorithm for large databases, a GPU/CPU hybrid architecture that exploits parallelism at all levels, and an optimized resource scheduler. We provide extensive experimental results demonstrating state-of-the-art performance in terms of recognition, scalability, and latency in real-world robotic applications.
[1] The size-frequency distributions of rocks >1.5 m diameter fully resolvable in High Resolution Imaging Science Experiment (HiRISE) images of the northern plains follow exponential models developed from lander measurements of smaller rocks and are continuous with rock distributions measured at the landing sites. Dark pixels at the resolution limit of Mars Orbiter Camera thought to be boulders are shown to be mostly dark shadows of clustered smaller rocks in HiRISE images. An automated rock detector algorithm that fits ellipses to shadows and cylinders to the rocks, accurately measured (within 1-2 pixels) rock diameter and height (by comparison to spacecraft of known size) of $10 million rocks over >1500 km 2 of the northern plains. Rock distributions in these counts parallel models for cumulative fractional area covered by 30-90% rocks in dense rock fields around craters, 10-30% rock coverage in less dense rock fields, and 0-10% rock coverage in background terrain away from craters. Above $1.5 m diameter, HiRISE resolves the same population of rocks seen in lander images, and thus size-frequency distributions can be extrapolated along model curves to estimate the number of rocks at smaller diameters. Extrapolating sparse rock distributions in the Phoenix landing ellipse indicate <1% chance of encountering a potentially hazardous rock during landing or that could impede the opening of the solar arrays. Extrapolations further suggest rocks large enough to depress the ground ice table and small enough to be picked up or pushed by the robotic arm should be present within reach for study after landing.
We present the discovery of KELT-21b, a hot Jupiter transiting the V=10.5 A8V star HD 332124. The planet has an orbital period of P=3.6127647±0.0000033 days and a radius of 1.586 0.040 0.039 -+ R J . We set an upper limit on the 1 planetary mass of M 3.91 P < M J at 3s confidence. We confirmed the planetary nature of the transiting companion using this mass limit and Doppler tomographic observations to verify that the companion transits HD 332124. These data also demonstrate that the planetary orbit is well-aligned with the stellar spin, with a sky-projected spinorbit misalignment of 5.6 1.91.7 . High-resolution imaging observations revealed the presence of a pair of stellar companions to KELT-21, located at a separation of 1 2 and with a combined contrast of K 6.39 0.06 S D = with respect to the primary. Although these companions are most likely physically associated with KELT-21, we cannot confirm this with our current data. If associated, the candidate companions KELT-21 B and C would each have masses of ∼0.12 M ☉ , a projected mutual separation of ∼20 au, and a projected separation of ∼500 au from KELT-21. KELT-21b may be one of only a handful of known transiting planets in hierarchical triple stellar systems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.