-
A search for the afterglows, kilonovae, and host galaxies of two short GRBs: GRB 211106A and GRB 211227A
Authors:
M. Ferro,
R. Brivio,
P. D'Avanzo,
A. Rossi,
L. Izzo,
S. Campana,
L. Christensen,
M. Dinatolo,
S. Hussein,
A. J. Levan,
A. Melandri,
M. G. Bernardini,
S. Covino,
V. D'Elia,
M. Della Valle,
M. De Pasquale,
B. P. Gompertz,
D. Hartmann,
K. E. Heintz,
P. Jakobsson,
C. Kouveliotou,
D. B. Malesani,
A. Martin-Carrillo,
L. Nava,
A. Nicuesa Guelbenzu
, et al. (8 additional authors not shown)
Abstract:
Context: GRB 211106A and GRB 211227A are recent gamma-ray bursts (GRBs) with initial X-ray positions suggesting associations with nearby galaxies (z < 0.7). Their prompt emission characteristics indicate GRB 211106A is a short-duration GRB and GRB 211227A is a short GRB with extended emission, likely originating from compact binary mergers. However, classifying solely based on prompt emission can…
▽ More
Context: GRB 211106A and GRB 211227A are recent gamma-ray bursts (GRBs) with initial X-ray positions suggesting associations with nearby galaxies (z < 0.7). Their prompt emission characteristics indicate GRB 211106A is a short-duration GRB and GRB 211227A is a short GRB with extended emission, likely originating from compact binary mergers. However, classifying solely based on prompt emission can be misleading. Aims: These short GRBs in the local Universe offer opportunities to search for associated kilonova (KN) emission and study host galaxy properties in detail. Methods: We conducted deep optical and NIR follow-up using ESO-VLT FORS2, HAWK-I, and MUSE for GRB 211106A, and ESO-VLT FORS2 and X-Shooter for GRB 211227A, starting shortly after the X-ray afterglow detection. We performed photometric analysis to look for afterglow and KN emissions associated with the bursts, along with host galaxy imaging and spectroscopy. Optical/NIR results were compared with Swift X-Ray Telescope (XRT) and other high-energy data. Results: For both GRBs we placed deep limits to the optical/NIR afterglow and KN emission. Host galaxies were identified: GRB 211106A at photometric z = 0.64 and GRB 211227A at spectroscopic z = 0.228. Host galaxy properties aligned with typical short GRB hosts. We also compared the properties of the bursts with the S-BAT4 sample to further examined the nature of these events. Conclusions: Study of prompt and afterglow phases, along with host galaxy analysis, confirms GRB 211106A as a short GRB and GRB 211227A as a short GRB with extended emission. The absence of optical/NIR counterparts is likely due to local extinction for GRB 211106A and a faint kilonova for GRB 211227A.
△ Less
Submitted 6 September, 2023;
originally announced September 2023.
-
The largest bright ULX population in a galaxy: X-ray variability and Luminosity Function in the Cartwheel ring Galaxy
Authors:
Chiara Salvaggio,
Anna Wolter,
Andrea Belfiore,
Monica Colpi
Abstract:
We analyse all the available Chandra observations of the Cartwheel Galaxy and its compact group, taken between 2001 and 2008, with the main aim of addressing the variability in the X-ray band for this spectacular collisional ring galaxy. We focus on the study of point-like sources, in particular we are interested in Ultraluminous X-ray sources (ULXs, Lx >= 10^39 erg/s), that we treat as a class. W…
▽ More
We analyse all the available Chandra observations of the Cartwheel Galaxy and its compact group, taken between 2001 and 2008, with the main aim of addressing the variability in the X-ray band for this spectacular collisional ring galaxy. We focus on the study of point-like sources, in particular we are interested in Ultraluminous X-ray sources (ULXs, Lx >= 10^39 erg/s), that we treat as a class. We exploit archival XMM-Newton data to enrich the study of the long-term variability, on timescales of months to years. We find a total of 44 sources in the group area, of which 37 in total are ULXs positionally linked with the galaxies and of which we can study variability. They are 29 in the Cartwheel itself, 7 in G1 and 1 in G3. About one third of these 37 sources show long-term variability, while no variability is detected within the single observations. Of those, 5 ULXs have a transient behaviour with a maximum range of variability (Lmax/Lmin) of about one order of magnitude and are the best candidate neutron stars. The X-ray Luminosity Function (XLF) of the point-like sources remains consistent in shape between the Chandra observations both for the Cartwheel galaxy itself and for G1, suggesting that flux variability does not strongly influence the average properties of the population on the observation timescales.
△ Less
Submitted 27 March, 2023;
originally announced March 2023.
-
A transient ultraluminous X-ray source in NGC 55
Authors:
A. Robba,
C. Pinto,
F. Pintore,
G. Rodriguez,
E. Ambrosi,
F. Barra,
G. Cusumano,
A. D'Aì,
M. Del Santo,
P. Kosec,
A. Marino,
M. Middleton,
T. Roberts,
C. Salvaggio,
R. Soria,
A. Wolter,
D. Walton
Abstract:
Ultraluminous X-ray sources (ULXs) are a class of accreting compact objects with X-ray luminosities above 10$^{39}$ erg s$^{-1}$. The average number of ULXs per galaxy is still not well constrained, especially given the uncertainty on the fraction of ULX transients. Here, we report the identification of a new transient ULX in the galaxy NGC 55 (which we label as ULX-2), thanks to recent XMM-Newton…
▽ More
Ultraluminous X-ray sources (ULXs) are a class of accreting compact objects with X-ray luminosities above 10$^{39}$ erg s$^{-1}$. The average number of ULXs per galaxy is still not well constrained, especially given the uncertainty on the fraction of ULX transients. Here, we report the identification of a new transient ULX in the galaxy NGC 55 (which we label as ULX-2), thanks to recent XMM-Newton and the Neil Gehrels Swift Observatory observations. This object was previously classified as a transient X-ray source with a luminosity around a few 10$^{38}$ erg s$^{-1}$ in a 2010 XMM-Newton observation. Thanks to new and deeper observations ($\sim$ 130 ks each), we show that the source reaches a luminosity peak $>1.6 \times 10^{39}$ erg s$^{-1}$. The X-ray spectrum of ULX-2 is much softer than in previous observations and fits in the class of soft ULXs. It can be well described using a model with two thermal components, as often found in ULXs. The time scales of the X-ray variability are of the order of a month and are likely driven by small changes in the accretion rate or due to super-orbital modulations, attributed to precession of the accretion disc, which is similar to other ULXs.
△ Less
Submitted 19 July, 2022;
originally announced July 2022.
-
Investigating the nature of the ultraluminous X-ray sources in the galaxy NGC 925
Authors:
Chiara Salvaggio,
Anna Wolter,
Fabio Pintore,
Ciro Pinto,
Elena Ambrosi,
Gian Luca Israel,
Alessio Marino,
Ruben Salvaterra,
Luca Zampieri,
Andrea Belfiore
Abstract:
Variability is a powerful tool to investigate properties of X-ray binaries (XRB), in particular for Ultraluminous X-ray sources (ULXs) that are mainly detected in the X-ray band. For most ULXs the nature of the accretor is unknown, although a few ULXs have been confirmed to be accreting at super-Eddington rates onto a neutron star (NS). Monitoring these sources is particularly useful both to detec…
▽ More
Variability is a powerful tool to investigate properties of X-ray binaries (XRB), in particular for Ultraluminous X-ray sources (ULXs) that are mainly detected in the X-ray band. For most ULXs the nature of the accretor is unknown, although a few ULXs have been confirmed to be accreting at super-Eddington rates onto a neutron star (NS). Monitoring these sources is particularly useful both to detect transients and to derive periodicities, linked to orbital and super-orbital modulations. Here we present the results of our monitoring campaign of the galaxy NGC 925, performed with the Neil Gehrels Swift Observatory. We also include archival and literature data obtained with Chandra, XMM-Newton and NuSTAR. We have studied spectra, light-curves and variability properties on days to months time-scales. All the three ULXs detected in this galaxy show flux variability. ULX-1 is one of the most luminous ULXs known, since only 10% of the ULXs exceed a luminosity of $\sim$5$\times$10$^{40}$ erg s$^{-1}$, but despite its high flux variability we found only weak spectral variability. We classify it as in a hard ultraluminous regime of super-Eddington accretion. ULX-2 and ULX-3 are less luminous but also variable in flux and possibly also in spectral shape. We classify them as in between the hard and the soft ultraluminous regimes. ULX-3 is a transient source: by applying a Lomb-Scargle algorithm we derive a periodicity of $\sim$ 126 d, which could be associated with an orbital or super-orbital origin.
△ Less
Submitted 22 February, 2022;
originally announced February 2022.
-
The rare X-ray flaring activity of the Ultraluminous X-ray source NGC 4559 X7
Authors:
Fabio Pintore,
S. Motta,
C. Pinto,
M. G. Bernardini,
G. Rodriguez-Castillo,
R. Salvaterra,
G. L. Israel,
P. Esposito,
E. Ambrosi,
C. Salvaggio,
L. Zampieri,
A. Wolter
Abstract:
Ultraluminous X-ray sources are considered amongst the most extremely accreting objects in the local Universe. The recent discoveries of pulsating neutron stars in ULXs strengthened the scenario of highly super-Eddington accretion mechanisms on stellar mass compact objects. In this work, we present the first long-term light curve of the source NGC 4559 X7 using all the available Swift, XMM-Newton,…
▽ More
Ultraluminous X-ray sources are considered amongst the most extremely accreting objects in the local Universe. The recent discoveries of pulsating neutron stars in ULXs strengthened the scenario of highly super-Eddington accretion mechanisms on stellar mass compact objects. In this work, we present the first long-term light curve of the source NGC 4559 X7 using all the available Swift, XMM-Newton, Chandra and NuSTAR data. Thanks to the high quality 2019 XMM-Newton and NuSTAR observations, we investigated in an unprecedented way the spectral and temporal properties of NGC 4559 X7. The source displayed flux variations of up to an order of magnitude and an unusual flaring activity. We modelled the spectra from NGC 4559 X7 with a combination of two thermal components, testing also the addition of a further high energy cut-off powerlaw. We observed a spectral hardening associated with a luminosity increase during the flares, and a spectral softening in the epochs far from the flares. Narrow absorption and emission lines were also found in the RGS spectra, suggesting the presence of an outflow. Furthermore, we measured hard and (weak) soft lags with magnitudes of a few hundreds of seconds whose origin is possibly be due to the accretion flow. We interpret the source properties in terms of a super-Eddington accretion scenario assuming the compact object is either a light stellar mass black hole or a neutron star.
△ Less
Submitted 26 March, 2021;
originally announced March 2021.
-
A Nearest Neighbor Network to Extract Digital Terrain Models from 3D Point Clouds
Authors:
Mohammed Yousefhussien,
David J. Kelbe,
Carl Salvaggio
Abstract:
When 3D-point clouds from overhead sensors are used as input to remote sensing data exploitation pipelines, a large amount of effort is devoted to data preparation. Among the multiple stages of the preprocessing chain, estimating the Digital Terrain Model (DTM) model is considered to be of a high importance; however, this remains a challenge, especially for raw point clouds derived from optical im…
▽ More
When 3D-point clouds from overhead sensors are used as input to remote sensing data exploitation pipelines, a large amount of effort is devoted to data preparation. Among the multiple stages of the preprocessing chain, estimating the Digital Terrain Model (DTM) model is considered to be of a high importance; however, this remains a challenge, especially for raw point clouds derived from optical imagery. Current algorithms estimate the ground points using either a set of geometrical rules that require tuning multiple parameters and human interaction, or cast the problem as a binary classification machine learning task where ground and non-ground classes are found. In contrast, here we present an algorithm that directly operates on 3D-point clouds and estimate the underlying DTM for the scene using an end-to-end approach without the need to classify points into ground and non-ground cover types. Our model learns neighborhood information and seamlessly integrates this with point-wise and block-wise global features. We validate our model using the ISPRS 3D Semantic Labeling Contest LiDAR data, as well as three scenes generated using dense stereo matching, representative of high-rise buildings, lower urban structures, and a dense old-city residential area. We compare our findings with two widely used software packages for DTM extraction, namely ENVI and LAStools. Our preliminary results show that the proposed method is able to achieve an overall Mean Absolute Error of 11.5% compared to 29% and 16% for ENVI and LAStools.
△ Less
Submitted 20 June, 2020; v1 submitted 21 May, 2020;
originally announced May 2020.
-
The Ultraluminous X-ray sources population of the galaxy NGC 7456
Authors:
F. Pintore,
M. Marelli,
R. Salvaterra,
G. L. Israel,
G. A. Rodríguez Castillo,
P. Esposito,
A. Belfiore,
A. De Luca,
A. Wolter,
S. Mereghetti,
L. Stella,
M. Rigoselli,
H. P. Earnshaw,
C. Pinto,
T. P. Roberts,
D. J. Walton,
F. Bernardini,
F. Haberl,
C. Salvaggio,
A. Tiengo,
L. Zampieri,
M. Bachetti,
M. Brightman,
P. Casella,
D. D'Agostino
, et al. (6 additional authors not shown)
Abstract:
Ultraluminous X-ray sources (ULXs) are a class of accreting compact objects with X-ray luminosities above 1e39 erg/s. The ULX population counts several hundreds objects but only a minor fraction is well studied. Here we present a detailed analysis of all ULXs hosted in the galaxy NGC 7456. It was observed in X-rays only once in the past (in 2005) by XMM-Newton, but the observation was short and st…
▽ More
Ultraluminous X-ray sources (ULXs) are a class of accreting compact objects with X-ray luminosities above 1e39 erg/s. The ULX population counts several hundreds objects but only a minor fraction is well studied. Here we present a detailed analysis of all ULXs hosted in the galaxy NGC 7456. It was observed in X-rays only once in the past (in 2005) by XMM-Newton, but the observation was short and strongly affected by high background. In 2018, we obtained a new, deeper (~90 ks) XMM-Newton observation that allowed us to perform a detailed characterization of the ULXs hosted in the galaxy. ULX-1 and ULX-2, the two brightest objects (Lx~(6-10)e39 erg/s), have spectra that can be described by a two-thermal component model as often found in ULXs. ULX-1 shows also one order of magnitude in flux variability on short-term timescales (hundreds to thousand ks). The other sources (ULX-3 and ULX-4) show flux changes of at least an order of magnitude, and these objects may be candidate transient ULXs although longer X-ray monitoring or further studies are required to ascribe them to the ULX population. In addition, we found a previously undetected source that might be a new candidate ULX (labelled as ULX-5) with a luminosity of ~1e39 erg/s and hard power-law spectral shape, whose nature is still unclear and for which a background Active Galactic Nucleus cannot be excluded. We discuss the properties of all the ULXs in NGC 7456 within the framework of super-Eddington accretion onto stellar mass compact objects. Although no pulsations were detected, we cannot exclude that the sources host neutron stars.
△ Less
Submitted 27 January, 2020; v1 submitted 23 January, 2020;
originally announced January 2020.
-
Diffuse X-ray emission around an ultraluminous X-ray pulsar
Authors:
Andrea Belfiore,
Paolo Esposito,
Fabio Pintore,
Giovanni Novara,
Ruben Salvaterra,
Andrea De Luca,
Andrea Tiengo,
Patrizia Caraveo,
Felix Fuerst,
Gian Luca Israel,
Danilo Magistrali,
Martino Marelli,
Sandro Mereghetti,
Alessandro Papitto,
Guillermo Rodriguez Castillo,
Chiara Salvaggio,
Luigi Stella,
Dominic Walton,
Anna Wolter,
Luca Zampieri
Abstract:
Ultraluminous X-ray sources (ULXs) are extragalactic X-ray emitters located off-center of their host galaxy and with a luminosity in excess of a few ${10^{39}\text{ erg s}^{-1}}$, if emitted isotropically. The discovery of periodic modulation revealed that in some ULXs the accreting compact object is a neutron star, indicating luminosities substantially above their Eddington limit. The most extrem…
▽ More
Ultraluminous X-ray sources (ULXs) are extragalactic X-ray emitters located off-center of their host galaxy and with a luminosity in excess of a few ${10^{39}\text{ erg s}^{-1}}$, if emitted isotropically. The discovery of periodic modulation revealed that in some ULXs the accreting compact object is a neutron star, indicating luminosities substantially above their Eddington limit. The most extreme object in this respect is ${NGC 5907~ULX-1}$ (ULX1), with a peak luminosity that is 500 times its Eddington limit. During a Chandra observation to probe a low state of ULX1, we detected diffuse X-ray emission at the position of ULX1. Its diameter is $2.7 \pm 1.0$ arcsec and contains 25 photons, none below 0.8 keV. We interpret this extended structure as an expanding nebula powered by the wind of ULX1. Its diameter of about ${200\text{ pc}}$, characteristic energy of ${\sim 1.9\text{ keV}}$, and luminosity of ${\sim 2\times10^{38}\text{ erg s}^{-1}}$ imply a mechanical power of ${1.3\times10^{41}\text{ erg s}^{-1}}$ and an age ${\sim 7 \times 10^{4}\text{ yr}}$. This interpretation suggests that a genuinely super-Eddington regime can be sustained for time scales much longer than the spin-up time of the neutron star powering the system. As the mechanical power from a single ULX nebula can rival the injection rate of cosmic rays of an entire galaxy, ULX nebulae could be important cosmic ray accelerators.
△ Less
Submitted 25 October, 2019;
originally announced October 2019.
-
Comparative study of panel and panelless-based reflectance conversion techniques for agricultural remote sensing
Authors:
Baabak Mamaghani,
Carl Salvaggio
Abstract:
Small unmanned aircraft systems (sUAS) have allowed for thousands of aerial images to be captured at a moments notice. The simplicity and relative low cost of flying a sUAS has provided remote sensing practitioners, both commercial and academic, with a viable alternative to traditional remote sensing platforms (airplanes and satellites). This paper is an expanded follow-up study to an initial work…
▽ More
Small unmanned aircraft systems (sUAS) have allowed for thousands of aerial images to be captured at a moments notice. The simplicity and relative low cost of flying a sUAS has provided remote sensing practitioners, both commercial and academic, with a viable alternative to traditional remote sensing platforms (airplanes and satellites). This paper is an expanded follow-up study to an initial work. Three radiance-to-reflectance conversion methods were tested to determine the optimal technique to use for converting raw digital count imagery to reflectance maps. The first two methods utilized in-scene reflectance conversion panels along with the empirical line method (ELM), while the final method used an upward looking sensor that recorded the band-effective spectral downwelling irradiance. The methods employed were 2-Point ELM, 1-point ELM, and At-altitude Radiance Ratio (AARR). The average signed reflectance factor errors produced by these methods on real sUAS imagery were: -0.005, -0.0028, and -0.0244 respectively. These errors were produced across four altitudes (150, 225, 300 and 375ft), six targets (grass, asphalt, concrete, blue felt, green felt and red felt), five spectral bands (blue, green, red, red edge and near infrared), and three weather conditions (cloudy, partly cloudy and sunny). Finally, data was simulated using the MODTRAN code to generate downwelling irradiance and sensor reaching radiance to compute the theoretical results of the AARR technique. A multitude of variables were varied for these simulations (atmosphere, time, day, target, sensor height, and visibility), which resulted in an overall theoretically achievable signed reflectance factor error of 0.0023.
△ Less
Submitted 8 October, 2019;
originally announced October 2019.
-
Batch-normalized Recurrent Highway Networks
Authors:
Chi Zhang,
Thang Nguyen,
Shagan Sah,
Raymond Ptucha,
Alexander Loui,
Carl Salvaggio
Abstract:
Gradient control plays an important role in feed-forward networks applied to various computer vision tasks. Previous work has shown that Recurrent Highway Networks minimize the problem of vanishing or exploding gradients. They achieve this by setting the eigenvalues of the temporal Jacobian to 1 across the time steps. In this work, batch normalized recurrent highway networks are proposed to contro…
▽ More
Gradient control plays an important role in feed-forward networks applied to various computer vision tasks. Previous work has shown that Recurrent Highway Networks minimize the problem of vanishing or exploding gradients. They achieve this by setting the eigenvalues of the temporal Jacobian to 1 across the time steps. In this work, batch normalized recurrent highway networks are proposed to control the gradient flow in an improved way for network convergence. Specifically, the introduced model can be formed by batch normalizing the inputs at each recurrence loop. The proposed model is tested on an image captioning task using MSCOCO dataset. Experimental results indicate that the batch normalized recurrent highway networks converge faster and performs better compared with the traditional LSTM and RHN based models.
△ Less
Submitted 26 September, 2018;
originally announced September 2018.
-
Semantic Sentence Embeddings for Paraphrasing and Text Summarization
Authors:
Chi Zhang,
Shagan Sah,
Thang Nguyen,
Dheeraj Peri,
Alexander Loui,
Carl Salvaggio,
Raymond Ptucha
Abstract:
This paper introduces a sentence to vector encoding framework suitable for advanced natural language processing. Our latent representation is shown to encode sentences with common semantic information with similar vector representations. The vector representation is extracted from an encoder-decoder model which is trained on sentence paraphrase pairs. We demonstrate the application of the sentence…
▽ More
This paper introduces a sentence to vector encoding framework suitable for advanced natural language processing. Our latent representation is shown to encode sentences with common semantic information with similar vector representations. The vector representation is extracted from an encoder-decoder model which is trained on sentence paraphrase pairs. We demonstrate the application of the sentence representations for two different tasks -- sentence paraphrasing and paragraph summarization, making it attractive for commonly used recurrent frameworks that process text. Experimental results help gain insight how vector representations are suitable for advanced language embedding.
△ Less
Submitted 26 September, 2018;
originally announced September 2018.
-
A new Ultraluminous X-ray source in the galaxy NGC 5907
Authors:
F. Pintore,
A. Belfiore,
G. Novara,
R. Salvaterra,
M. Marelli,
A. De Luca,
M. Rigoselli,
G. Israel,
G. Rodriguez,
S. Mereghetti,
A. Wolter,
D. J. Walton,
F. Fuerst,
E. Ambrosi,
L. Zampieri,
A. Tiengo,
C. Salvaggio
Abstract:
We report on the serendipitous discovery of a new transient in NGC 5907, at a peak luminosity of 6.4x10^{39} erg/s. The source was undetected in previous 2012 Chandra observations with a 3 sigma upper limit on the luminosity of 1.5x10^{38} erg/s, implying a flux increase of a factor of >35. We analyzed three recent 60ks/50ks Chandra and 50ks XMM-Newton observations, as well as all the available Sw…
▽ More
We report on the serendipitous discovery of a new transient in NGC 5907, at a peak luminosity of 6.4x10^{39} erg/s. The source was undetected in previous 2012 Chandra observations with a 3 sigma upper limit on the luminosity of 1.5x10^{38} erg/s, implying a flux increase of a factor of >35. We analyzed three recent 60ks/50ks Chandra and 50ks XMM-Newton observations, as well as all the available Swift observations performed between August 2017/March 2018. Until the first half of October 2017, Swift observations do not show any emission from the source. The transient entered the ULX regime in less than two weeks and its outburst was still on-going at the end of February 2018. The 0.3-10 keV spectrum is consistent with a single multicolour blackbody disc (kT~1.5 keV). The source might be a ~30 solar mass black hole accreting at the Eddington limit. However, although we did not find evidence of pulsations, we cannot rule-out the possibility that this ULX hosts an accreting neutron star.
△ Less
Submitted 10 June, 2018;
originally announced June 2018.
-
An initial exploration of vicarious and in-scene calibration techniques for small unmanned aircraft systems
Authors:
Baabak Mamaghani,
Geoffrey Sasaki,
Ryan Connal,
Kevin Kha,
Jackson Knappen,
Ryan Hartzell,
Evan Marcellus,
Timothy Bauch,
Nina Raqueno,
Carl Salvaggio
Abstract:
The use of small unmanned aircraft systems (sUAS) for applications in the field of precision agriculture has demonstrated the need to produce temporally consistent imagery to allow for quantitative comparisons. In order for these aerial images to be used to identify actual changes on the ground, conversion of raw digital count to reflectance, or to an atmospherically normalized space, needs to be…
▽ More
The use of small unmanned aircraft systems (sUAS) for applications in the field of precision agriculture has demonstrated the need to produce temporally consistent imagery to allow for quantitative comparisons. In order for these aerial images to be used to identify actual changes on the ground, conversion of raw digital count to reflectance, or to an atmospherically normalized space, needs to be carried out. This paper will describe an experiment that compares the use of reflectance calibration panels, for use with the empirical line method (ELM), against a newly proposed ratio of the target radiance and the downwelling radiance, to predict the reflectance of known targets in the scene. We propose that the use of an on-board downwelling light sensor (DLS) may provide the sUAS remote sensing practitioner with an approach that does not require the expensive and time consuming task of placing known reflectance standards in the scene. Three calibration methods were tested in this study: 2-Point ELM, 1-Point ELM, and At-altitude Radiance Ratio (AARR). Our study indicates that the traditional 2-Point ELM produces the lowest mean error in band effective reflectance factor, 0.0165. The 1-Point ELM and AARR produce mean errors of 0.0343 and 0.0287 respectively. A modeling of the proposed AARR approach indicates that the technique has the potential to perform better than the 2-Point ELM method, with a 0.0026 mean error in band effective reflectance factor, indicating that this newly proposed technique may prove to be a viable alternative with suitable on-board sensors.
△ Less
Submitted 25 April, 2018;
originally announced April 2018.
-
A Fully Convolutional Network for Semantic Labeling of 3D Point Clouds
Authors:
Mohammed Yousefhussien,
David J. Kelbe,
Emmett J. Ientilucci,
Carl Salvaggio
Abstract:
When classifying point clouds, a large amount of time is devoted to the process of engineering a reliable set of features which are then passed to a classifier of choice. Generally, such features - usually derived from the 3D-covariance matrix - are computed using the surrounding neighborhood of points. While these features capture local information, the process is usually time-consuming, and requ…
▽ More
When classifying point clouds, a large amount of time is devoted to the process of engineering a reliable set of features which are then passed to a classifier of choice. Generally, such features - usually derived from the 3D-covariance matrix - are computed using the surrounding neighborhood of points. While these features capture local information, the process is usually time-consuming, and requires the application at multiple scales combined with contextual methods in order to adequately describe the diversity of objects within a scene. In this paper we present a 1D-fully convolutional network that consumes terrain-normalized points directly with the corresponding spectral data,if available, to generate point-wise labeling while implicitly learning contextual features in an end-to-end fashion. Our method uses only the 3D-coordinates and three corresponding spectral features for each point. Spectral features may either be extracted from 2D-georeferenced images, as shown here for Light Detection and Ranging (LiDAR) point clouds, or extracted directly for passive-derived point clouds,i.e. from muliple-view imagery. We train our network by splitting the data into square regions, and use a pooling layer that respects the permutation-invariance of the input points. Evaluated using the ISPRS 3D Semantic Labeling Contest, our method scored second place with an overall accuracy of 81.6%. We ranked third place with a mean F1-score of 63.32%, surpassing the F1-score of the method with highest accuracy by 1.69%. In addition to labeling 3D-point clouds, we also show that our method can be easily extended to 2D-semantic segmentation tasks, with promising initial results.
△ Less
Submitted 3 October, 2017;
originally announced October 2017.
-
Algorithms for Semantic Segmentation of Multispectral Remote Sensing Imagery using Deep Learning
Authors:
Ronald Kemker,
Carl Salvaggio,
Christopher Kanan
Abstract:
Deep convolutional neural networks (DCNNs) have been used to achieve state-of-the-art performance on many computer vision tasks (e.g., object recognition, object detection, semantic segmentation) thanks to a large repository of annotated image data. Large labeled datasets for other sensor modalities, e.g., multispectral imagery (MSI), are not available due to the large cost and manpower required.…
▽ More
Deep convolutional neural networks (DCNNs) have been used to achieve state-of-the-art performance on many computer vision tasks (e.g., object recognition, object detection, semantic segmentation) thanks to a large repository of annotated image data. Large labeled datasets for other sensor modalities, e.g., multispectral imagery (MSI), are not available due to the large cost and manpower required. In this paper, we adapt state-of-the-art DCNN frameworks in computer vision for semantic segmentation for MSI imagery. To overcome label scarcity for MSI data, we substitute real MSI for generated synthetic MSI in order to initialize a DCNN framework. We evaluate our network initialization scheme on the new RIT-18 dataset that we present in this paper. This dataset contains very-high resolution MSI collected by an unmanned aircraft system. The models initialized with synthetic imagery were less prone to over-fitting and provide a state-of-the-art baseline for future work.
△ Less
Submitted 1 May, 2018; v1 submitted 19 March, 2017;
originally announced March 2017.
-
High-Resolution Multispectral Dataset for Semantic Segmentation
Authors:
Ronald Kemker,
Carl Salvaggio,
Christopher Kanan
Abstract:
Unmanned aircraft have decreased the cost required to collect remote sensing imagery, which has enabled researchers to collect high-spatial resolution data from multiple sensor modalities more frequently and easily. The increase in data will push the need for semantic segmentation frameworks that are able to classify non-RGB imagery, but this type of algorithmic development requires an increase in…
▽ More
Unmanned aircraft have decreased the cost required to collect remote sensing imagery, which has enabled researchers to collect high-spatial resolution data from multiple sensor modalities more frequently and easily. The increase in data will push the need for semantic segmentation frameworks that are able to classify non-RGB imagery, but this type of algorithmic development requires an increase in publicly available benchmark datasets with class labels. In this paper, we introduce a high-resolution multispectral dataset with image labels. This new benchmark dataset has been pre-split into training/testing folds in order to standardize evaluation and continue to push state-of-the-art classification frameworks for non-RGB imagery.
△ Less
Submitted 6 March, 2017;
originally announced March 2017.
