Towards machine learning assisted error identification in orbital laser altimetry for tides derivation

2021 ◽  
Author(s):  
Oliver Stenzel ◽  
Martin Hilchenbach
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Training Dataset ◽  
Altimeter Data ◽  
Laser Altimetry ◽  
Linear Network ◽  
Error Identification ◽  
Laser Altimeter ◽  
Convolutional Network ◽  
The Impact

<p>Laser altimetry experiments on the NASA MESSENGER mission [1], and on the currently on cruise ESA/JAXA BepiColombo Mission [2,3] did and are going to yield, respectively, a plethora of range measurements of the surface of Mercury. Orbital laser altimetry can be used to derive tidal parameters, which can in turn be used to infer properties of a body’s interior [4,5]. The derivation of tidal parameters requires large datasets of precise and accurate measurements. Errors as well as outliers can degrade the quality of the computed tidal parameters. While many outliers can be filtered though conventional automated processes, other errors could only be identified by human supervision. In the face of the amount of data involved, systematic user interaction at the error identification step becomes unpractical. A neural network trained with user expertise could help spotting outliers and errors and would improve the derived parameters in accuracy and precession. We started developing a neural network based on the pytorch framework[6] and compared the performance with a small training dataset form the MESSENGER Laser Altimeter (MLA) for a linear and a convolutional network. The results were much in favour of the linear network [7]. In this presentation we explore the reasons behind bad convolutional network performance with extended training and test datasets. We are going to show our results for the filtered datasets and the impact this has on the derived tidal parameters. The filtering with an artificial neural network might be useful for other applications, as well.</p> <p>1. Cavanaugh, J. F. et al. The Mercury Laser Altimeter Instrument for the MESSENGER Mission. Space Sci Rev 131, 451–479 (2007).</p> <p>2. Benkhoff, J. et al. BepiColombo—Comprehensive exploration of Mercury: Mission overview and science goals. Planetary and Space Science 58, 2–20 (2010).</p> <p>3. Thomas, N. et al. The BepiColombo Laser Altimeter. Space Sci Rev 217, 25 (2021).</p> <p>4. Thor, R. N. et al. Determination of the lunar body tide from global laser altimetry data. J Geod 95, 4 (2021).</p> <p>5. Thor, R. N. et al. Prospects for measuring Mercury’s tidal Love number h2 with the BepiColombo Laser Altimeter. A&A 633, A85 (2020).</p> <p>6. Paszke, A., et al., PyTorch: An Imperative Style, High-Performance Deep Learning Library, In: Advances in Neural Information Processing Systems 32, pp 8024–8035, 2019.</p> <p>7. Stenzel, O., Thor, R., and Hilchenbach, M.: Error identification in orbital laser altimeter data by machine learning, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14749, https://doi.org/10.5194/egusphere-egu21-14749, 2021.</p>

Error identification in orbital laser altimeter data by machine learning

2021 ◽  
Author(s):  
Oliver Stenzel ◽  
Robin Thor ◽  
Martin Hilchenbach
Keyword(s):  
Machine Learning ◽  
High Performance ◽  
Space Science ◽  
Altimeter Data ◽  
Laser Altimetry ◽  
Error Identification ◽  
Laser Altimeter ◽  
Data Set ◽  
Baseline Design ◽  
Solar System Bodies

<p>Orbital Laser altimeters deliver a plethora of data that is used to map planetary surfaces [1] and to understand interiors of solar system bodies [2]. Accuracy and precision of laser altimetry measurements depend on the knowledge of spacecraft position and pointing and on the instrument. Both are important for the retrieval of tidal parameters. In order to assess the quality of the altimeter retrievals, we are training and implementing an artificial neural network (ANN) to identify and exclude scans from analysis which yield erroneous data. The implementation is based on the PyTorch framework [3]. We are presenting our results for the MESSENGER Mercury Laser Altimeter (MLA) data set [4], but also in view of future analysis of the BepiColombo Laser Altimeter (BELA) data, which will arrive in orbit around Mercury in 2025 on board the Mercury Planetary Orbiter [5,6]. We further explore conventional methods of error identification and compare these with the machine learning results. Short periods of large residuals or large variation of residuals are identified and used to detect erroneous measurements. Furthermore, long-period systematics, such as those caused by slow variations in instrument pointing, can be modelled by including additional parameters.<br>[1] Zuber, Maria T., David E. Smith, Roger J. Phillips, Sean C. Solomon, Gregory A. Neumann, Steven A. Hauck, Stanton J. Peale, et al. ‘Topography of the Northern Hemisphere of Mercury from MESSENGER Laser Altimetry’. Science 336, no. 6078 (13 April 2012): 217–20. https://doi.org/10.1126/science.1218805.<br>[2] Thor, Robin N., Reinald Kallenbach, Ulrich R. Christensen, Philipp Gläser, Alexander Stark, Gregor Steinbrügge, and Jürgen Oberst. ‘Determination of the Lunar Body Tide from Global Laser Altimetry Data’. Journal of Geodesy 95, no. 1 (23 December 2020): 4. https://doi.org/10.1007/s00190-020-01455-8.<br>[3] Paszke, Adam, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, et al. ‘PyTorch: An Imperative Style, High-Performance Deep Learning Library’. Advances in Neural Information Processing Systems 32 (2019): 8026–37.<br>[4] Cavanaugh, John F., James C. Smith, Xiaoli Sun, Arlin E. Bartels, Luis Ramos-Izquierdo, Danny J. Krebs, Jan F. McGarry, et al. ‘The Mercury Laser Altimeter Instrument for the MESSENGER Mission’. Space Science Reviews 131, no. 1 (1 August 2007): 451–79. https://doi.org/10.1007/s11214-007-9273-4.<br>[5] Thomas, N., T. Spohn, J. -P. Barriot, W. Benz, G. Beutler, U. Christensen, V. Dehant, et al. ‘The BepiColombo Laser Altimeter (BELA): Concept and Baseline Design’. Planetary and Space Science 55, no. 10 (1 July 2007): 1398–1413. https://doi.org/10.1016/j.pss.2007.03.003.<br>[6] Benkhoff, Johannes, Jan van Casteren, Hajime Hayakawa, Masaki Fujimoto, Harri Laakso, Mauro Novara, Paolo Ferri, Helen R. Middleton, and Ruth Ziethe. ‘BepiColombo—Comprehensive Exploration of Mercury: Mission Overview and Science Goals’. Planetary and Space Science, Comprehensive Science Investigations of Mercury: The scientific goals of the joint ESA/JAXA mission BepiColombo, 58, no. 1 (1 January 2010): 2–20. https://doi.org/10.1016/j.pss.2009.09.020.</p>

scholarly journals Encoding Health Records into Pathway Representations for Deep Learning

2021 ◽  
Author(s):  
Marco Luca Sbodio ◽  
Natasha Mulligan ◽  
Stefanie Speichert ◽  
Vanessa Lopez ◽  
Joao Bettencourt-Silva
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Deep Learning ◽  
Source Code ◽  
Training Dataset ◽  
Health Records ◽  
Learning Tasks ◽  
Patient Pathways ◽  
Computational Resources ◽  
The Impact

There is a growing trend in building deep learning patient representations from health records to obtain a comprehensive view of a patient’s data for machine learning tasks. This paper proposes a reproducible approach to generate patient pathways from health records and to transform them into a machine-processable image-like structure useful for deep learning tasks. Based on this approach, we generated over a million pathways from FAIR synthetic health records and used them to train a convolutional neural network. Our initial experiments show the accuracy of the CNN on a prediction task is comparable or better than other autoencoders trained on the same data, while requiring significantly less computational resources for training. We also assess the impact of the size of the training dataset on autoencoders performances. The source code for generating pathways from health records is provided as open source.

scholarly journals Evaluation of Power Insulator Detection Efficiency with the Use of Limited Training Dataset

2020 ◽  
Vol 10 (6) ◽  
pp. 2104
Author(s):  
Michał Tomaszewski ◽  
Paweł Michalski ◽  
Jakub Osuchowski
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Object Detection ◽  
Convolutional Neural Network ◽  
Deep Neural Networks ◽  
Detection Efficiency ◽  
Training Data ◽  
Training Dataset ◽  
Training Set ◽  
Convolutional Network

This article presents an analysis of the effectiveness of object detection in digital images with the application of a limited quantity of input. The possibility of using a limited set of learning data was achieved by developing a detailed scenario of the task, which strictly defined the conditions of detector operation in the considered case of a convolutional neural network. The described solution utilizes known architectures of deep neural networks in the process of learning and object detection. The article presents comparisons of results from detecting the most popular deep neural networks while maintaining a limited training set composed of a specific number of selected images from diagnostic video. The analyzed input material was recorded during an inspection flight conducted along high-voltage lines. The object detector was built for a power insulator. The main contribution of the presented papier is the evidence that a limited training set (in our case, just 60 training frames) could be used for object detection, assuming an outdoor scenario with low variability of environmental conditions. The decision of which network will generate the best result for such a limited training set is not a trivial task. Conducted research suggests that the deep neural networks will achieve different levels of effectiveness depending on the amount of training data. The most beneficial results were obtained for two convolutional neural networks: the faster region-convolutional neural network (faster R-CNN) and the region-based fully convolutional network (R-FCN). Faster R-CNN reached the highest AP (average precision) at a level of 0.8 for 60 frames. The R-FCN model gained a worse AP result; however, it can be noted that the relationship between the number of input samples and the obtained results has a significantly lower influence than in the case of other CNN models, which, in the authors’ assessment, is a desired feature in the case of a limited training set.

scholarly journals Applying of machine learning in the construction of a voice-controlled interface on the example of a music player

2019 ◽  
Vol 13 ◽  
pp. 302-309
Author(s):  
Jakub Basiakowski
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Convolutional Neural Network ◽  
Feedforward Neural Network ◽  
Hidden Layer ◽  
Music Player ◽  
The Impact

The following paper presents the results of research on the impact of machine learning in the construction of a voice-controlled interface. Two different models were used for the analysys: a feedforward neural network containing one hidden layer and a more complicated convolutional neural network. What is more, a comparison of the applied models was presented. This comparison was performed in terms of quality and the course of training.

Thermal–Mechanical Coupling Evaluation of the Panel Performance of a Prefabricated Cabin-Type Substation Based on Machine Learning

Fire ◽  
2021 ◽  
Vol 4 (4) ◽  
pp. 93
Author(s):  
Xiangsheng Lei ◽  
Jinwu Ouyang ◽  
Yanfeng Wang ◽  
Xinghua Wang ◽  
Xiaofeng Zhang ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Numerical Simulation ◽  
Fire Resistance ◽  
Bp Neural Network ◽  
Stress Test ◽  
Normal Operation ◽  
Model Parameters ◽  
Mechanical Coupling ◽  
The Impact

The panel performance of a prefabricated cabin-type substation under the impact of fires plays a vital role in the normal operation of the substation. However, current evaluations of the panel performance of substations under fire still focus on fire resistance tests, which seldom consider the relationship between fire behavior and the mechanical load of the panel under the impact of fires. Aiming at the complex and uncertain relationship between the thermal and mechanical performance of the substation panel under impact of fires, this paper proposes a machine learning method based on a BP neural network. First, the fire resistance test and the stress test of the panel is carried out, then a machine learning model is established based on the BP neural network. According to the collected data, the model parameters are obtained through a series of training and verification processes. Meanwhile, the correlation between the panel performance and fire resistance was obtained. Finally, related parameters are input into the thermal–mechanical coupling evaluation model for the substation panel performance to evaluate the fire resistance performance of the substation panel. To verify the correctness of the established model, numerical simulation of the fire test and stress test of the panel is conducted, and numerical simulation samples are predicted by the trained model. The results show that the prediction curve of neural network is closer to the real results compared with the numerical simulation, and the established model can accurately evaluate the thermal–mechanical coupling performance of the substation panel under fire.

The impact of Negative to Positive Training Dataset Ratio on Atrial Fibrillation Classification Machine Learning Algorithms Performance

2020 ◽  
Vol 1500 ◽  
pp. 012131
Author(s):  
Firdaus ◽  
Andre Herviant Juliano ◽  
Naufal Rachmatullah ◽  
Sarifah Putri Rafflesia ◽  
Dinna Yunika Hardiyanti ◽  
...  
Keyword(s):  
Machine Learning ◽  
Atrial Fibrillation ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Training Dataset ◽  
The Impact

scholarly journals Interannual changes of the floating ice shelf of Petermann Gletscher, North Greenland, from 2000 to 2012

2014 ◽  
Vol 60 (221) ◽  
pp. 489-499 ◽  
Author(s):  
Andreas Münchow ◽  
Laurie Padman ◽  
Helen A. Fricker
Keyword(s):  
Mass Loss ◽  
Three Dimensional ◽  
Greenland Ice Sheet ◽  
Altimeter Data ◽  
Ice Shelf ◽  
Laser Altimetry ◽  
Laser Altimeter ◽  
Airborne Laser ◽  
Tracking Surface ◽  
Basal Melting

AbstractPetermann Gletscher, northwest Greenland, drains 4% of the Greenland ice sheet into Nares Strait. Its floating ice shelf retreated from 81 to 48 km in length during two large calving events in 2010 and 2012. We document changes in the three-dimensional ice-shelf structure from 2000 to 2012, using repeated tracks of airborne laser altimetry and ice radio-echo sounding, ICESat laser altimetry and MODIS visible imagery. The recent ice-shelf velocity, measured by tracking surface features between flights in 2010 and 2011, is ~1.25 km a−1, ~15–30% faster than estimates made before 2010. The steady- state along-flow ice divergence represents 6.3 Gta−1 mass loss through basal melting (~5Gta−1) and surface melting and sublimation (~1.0Gta−1). Airborne laser altimeter data reveal thinning, both along a thin central channel and on the thicker ambient ice shelf. From 2007 to 2010 the ice shelf thinned by ~5 m a−1, which represents a non-steady mass loss of ~4.1 Gta−1. We suggest that thinning in the basal channels structurally weakened the ice shelf and may have played a role in the recent calving events.

scholarly journals IMPROVE THE ZY-3 HEIGHT ACCURACY USING ICESAT/GLAS LASER ALTIMETER DATA

2016 ◽  
Vol XLI-B1 ◽  
pp. 37-42
Author(s):  
Guoyuan Li ◽  
Xinming Tang ◽  
Xiaoming Gao ◽  
Chongyang Zhang ◽  
Tao Li
Keyword(s):  
High Resolution ◽  
Bundle Adjustment ◽  
Experimental Result ◽  
Altimeter Data ◽  
Surface Model ◽  
Laser Altimetry ◽  
Laser Altimeter ◽  
Ice Cloud ◽  
Elevation Data ◽  
Polar Ice Sheets

ZY-3 is the first civilian high resolution stereo mapping satellite, which has been launched on 9th, Jan, 2012. The aim of ZY-3 satellite is to obtain high resolution stereo images and support the 1:50000 scale national surveying and mapping. Although ZY-3 has very high accuracy for direct geo-locations without GCPs (Ground Control Points), use of some GCPs is still indispensible for high precise stereo mapping. The GLAS (Geo-science Laser Altimetry System) loaded on the ICESat (Ice Cloud and land Elevation Satellite), which is the first laser altimetry satellite for earth observation. GLAS has played an important role in the monitoring of polar ice sheets, the measuring of land topography and vegetation canopy heights after launched in 2003. Although GLAS has ended in 2009, the derived elevation dataset still can be used after selection by some criteria. <br><br> In this paper, the ICESat/GLAS laser altimeter data is used as height reference data to improve the ZY-3 height accuracy. A selection method is proposed to obtain high precision GLAS elevation data. Two strategies to improve the ZY-3 height accuracy are introduced. One is the conventional bundle adjustment based on RFM and bias-compensated model, in which the GLAS footprint data is viewed as height control. The second is to correct the DSM (Digital Surface Model) straightly by simple block adjustment, and the DSM is derived from the ZY-3 stereo imaging after freedom adjustment and dense image matching. The experimental result demonstrates that the height accuracy of ZY-3 without other GCPs can be improved to 3.0 meter after adding GLAS elevation data. What’s more, the comparison of the accuracy and efficiency between the two strategies is implemented for application.

scholarly journals Predicting Alert Source Device using Machine Learning Algorithms

2020 ◽  
Vol 9 (9) ◽  
pp. 1-10
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Model Building ◽  
Learning Algorithm ◽  
Learning Algorithms ◽  
Research Work ◽  
Machine Learning Algorithms ◽  
Training Dataset ◽  
Imbalanced Dataset ◽  
Daunting Task

In a large distributed virtualized environment, predicting the alerting source from its text seems to be daunting task. This paper explores the option of using machine learning algorithm to solve this problem. Unfortunately, our training dataset is highly imbalanced. Where 96% of alerting data is reported by 24% of alerting sources. This is the expected dataset in any live distributed virtualized environment, where new version of device will have relatively less alert compared to older devices. Any classification effort with such imbalanced dataset present different set of challenges compared to binary classification. This type of skewed data distribution makes conventional machine learning less effective, especially while predicting the minority device type alerts. Our challenge is to build a robust model which can cope with this imbalanced dataset and achieves relative high level of prediction accuracy. This research work stared with traditional regression and classification algorithms using bag of words model. Then word2vec and doc2vec models are used to represent the words in vector formats, which preserve the sematic meaning of the sentence. With this alerting text with similar message will have same vector form representation. This vectorized alerting text is used with Logistic Regression for model building. This yields better accuracy, but the model is relatively complex and demand more computational resources. Finally, simple neural network is used for this multi-class text classification problem domain by using keras and tensorflow libraries. A simple two layered neural network yielded 99 % accuracy, even though our training dataset was not balanced. This paper goes through the qualitative evaluation of the different machine learning algorithms and their respective result. Finally, two layered deep learning algorithms is selected as final solution, since it takes relatively less resource and time with better accuracy values.

scholarly journals Evaluating structural safety of trusses using Machine Learning

2021 ◽  
Vol 15 (58) ◽  
pp. 308-318
Author(s):  
Tran-Hieu Nguyen ◽  
Anh-Tuan Vu
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Support Vector Machine ◽  
Roc Curve ◽  
Deep Neural Network ◽  
Structural Safety ◽  
Training Dataset ◽  
Support Vector ◽  
Machine Model ◽  
Adaptive Boosting

In this paper, a machine learning-based framework is developed to quickly evaluate the structural safety of trusses. Three numerical examples of a 10-bar truss, a 25-bar truss, and a 47-bar truss are used to illustrate the proposed framework. Firstly, several truss cases with different cross-sectional areas are generated by employing the Latin Hypercube Sampling method. Stresses inside truss members as well as displacements of nodes are determined through finite element analyses and obtained values are compared with design constraints. According to the constraint verification, the safety state is assigned as safe or unsafe. Members’ sectional areas and the safety state are stored as the inputs and outputs of the training dataset, respectively. Three popular machine learning classifiers including Support Vector Machine, Deep Neural Network, and Adaptive Boosting are used for evaluating the safety of structures. The comparison is conducted based on two metrics: the accuracy and the area under the ROC curve. For the two first examples, three classifiers get more than 90% of accuracy. For the 47-bar truss, the accuracies of the Support Vector Machine model and the Deep Neural Network model are lower than 70% but the Adaptive Boosting model still retains the high accuracy of approximately 98%. In terms of the area under the ROC curve, the comparative results are similar. Overall, the Adaptive Boosting model outperforms the remaining models. In addition, an investigation is carried out to show the influence of the parameters on the performance of the Adaptive Boosting model.

Sign in / Sign up

Export Citation Format

Share Document