In this paper, we consider convolutional neural networks operating on sparse inputs with an application to depth upsampling from sparse laser scan data. First, we show that traditional convolutional networks perform poorly when applied to sparse data even when the location of missing data is provided to the network. To overcome this problem, we propose a simple yet effective sparse convolution layer which explicitly considers the location of missing data during the convolution operation. We demonstrate the benefits of the proposed network architecture in synthetic and real experiments with respect to various baseline approaches. Compared to dense baselines, the proposed sparse convolution network generalizes well to novel datasets and is invariant to the level of sparsity in the data. For our evaluation, we derive a novel dataset from the KITTI benchmark, comprising 93k depth annotated RGB images. Our dataset allows for training and evaluating depth upsampling and depth prediction techniques in challenging real-world settings and will be made available upon publication.
Results from the occultation of the sun by Neptune imply a temperature of 750 +/- 150 kelvins in the upper levels of the atmosphere (composed mostly of atomic and molecular hydrogen) and define the distributions of methane, acetylene, and ethane at lower levels. The ultraviolet spectrum of the sunlit atmosphere of Neptune resembles the spectra of the Jupiter, Saturn, and Uranus atmospheres in that it is dominated by the emissions of H Lyman alpha (340 +/- 20 rayleighs) and molecular hydrogen. The extreme ultraviolet emissions in the range from 800 to 1100 angstroms at the four planets visited by Voyager scale approximately as the inverse square of their heliocentric distances. Weak auroral emissions have been tentatively identified on the night side of Neptune. Airglow and occultation observations of Triton's atmosphere show that it is composed mainly of molecular nitrogen, with a trace of methane near the surface. The temperature of Triton's upper atmosphere is 95 +/- 5 kelvins, and the surface pressure is roughly 14 microbars.
An impressive amount of evidence supports the proposal of Alvarez et al. that the Cretaceous era was ended abruptly by the impact of a comet or asteroid. The recent discovery of an apparently global soot layer at the Cretaceous/Tertiary boundary indicates that global wildfires were somehow ignited by the impact. Here we show that the thermal radiation produced by the ballistic re-entry of ejecta condensed from the vapour plume of the impact could have increased the global radiation flux by factors of 50 to 150 times the solar input for periods ranging from one to several hours. This great increase in thermal radiation may have been responsible for the ignition of global wildfires, as well as having deleterious effects on unprotected animal life.
Abstract-In this paper, we present RegNet, the first deep convolutional neural network (CNN) to infer a 6 degrees of freedom (DOF) extrinsic calibration between multimodal sensors, exemplified using a scanning LiDAR and a monocular camera. Compared to existing approaches, RegNet casts all three conventional calibration steps (feature extraction, feature matching and global regression) into a single real-time capable CNN. Our method does not require any human interaction and bridges the gap between classical offline and target-less online calibration approaches as it provides both a stable initial estimation as well as a continuous online correction of the extrinsic parameters. During training we randomly decalibrate our system in order to train RegNet to infer the correspondence between projected depth measurements and RGB image and finally regress the extrinsic calibration. Additionally, with an iterative execution of multiple CNNs, that are trained on different magnitudes of decalibration, our approach compares favorably to state-of-the-art methods in terms of a mean calibration error of 0.28• for the rotational and 6 cm for the translation components even for large decalibrations up to 1.5 m and 20• .
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.