SEARCH · Search NASA

Results for “Computer Vision”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 199 records · Page 11

All sources NASA NTRS DOE OSTI

NASA NTRS2016

UAV Research at NASA Langley: Towards Safe, Reliable, and Autonomous Operations

Unmanned Aerial Vehicles (UAV) are fundamental components in several aspects of research at NASA Langley, such as flight dynamics, mission-driven airframe design, airspace integration demonstrations, atmospheric science projects, and more. In particular, NASA Langley Research Center (Langley) is using UAVs to develop and demonstrate innovative capabilities that meet the autonomy and robotics challenges that are anticipated in science, space exploration, and aeronautics. These capabilities will enable new NASA missions such as asteroid rendezvous and retrieval (ARRM), Mars exploration, in-situ resource utilization (ISRU), pollution measurements in historically inaccessible areas, and the integration of UAVs into our everyday lives all missions of increasing complexity, distance, pace, and/or accessibility. Building on decades of NASA experience and success in the design, fabrication, and integration of robust and reliable automated systems for space and aeronautics, Langley Autonomy Incubator seeks to bridge the gap between automation and autonomy by enabling safe autonomous operations via onboard sensing and perception systems in both data-rich and data-deprived environments. The Autonomy Incubator is focused on the challenge of mobility and manipulation in dynamic and unstructured environments by integrating technologies such as computer vision, visual odometry, real-time mapping, path planning, object detection and avoidance, object classification, adaptive control, sensor fusion, machine learning, and natural human-machine teaming. These technologies are implemented in an architectural framework developed in-house for easy integration and interoperability of cutting-edge hardware and software.

Davila, Carlos G.↗

NASA NTRS2017

Detection, Localization, and Tracking of Unauthorized UAS and Jammers

Small unmanned aircraft systems (UASs) are expected to take major roles in future smart cities, for example, by delivering goods and merchandise, potentially serving as mobile hot spots for broadband wireless access, and maintaining surveillance and security. Although they can be used for the betterment of the society, they can also be used by malicious entities to conduct physical and cyber attacks to infrastructure, private/public property, and people. Even for legitimate use-cases of small UASs, air traffic management (ATM) for UASs becomes of critical importance for maintaining safe and collusion-free operation. Therefore, various ways to detect, track, and interdict potentially unauthorized drones carries critical importance for surveillance and ATM applications. In this paper, we will review techniques that rely on ambient radio frequency signals (emitted from UASs), radars, acoustic sensors, and computer vision techniques for detection of malicious UASs. We will present some early experimental and simulation results on radar-based range estimation of UASs, and receding horizon tracking of UASs. Subsequently, we will overview common techniques that are considered for interdiction of UASs.

surveillance↗

NASA NTRS2017

Detection of Hail Storms in Radar Imagery Using Deep Learning

In 2016, hail was responsible for 3.5 billion and 23 million dollars in damage to property and crops, respectively, making it the second costliest weather phenomenon in the United States. In an effort to improve hail-prediction techniques and reduce the societal impacts associated with hail storms, we propose a deep learning technique that leverages radar imagery for automatic detection of hail storms. The technique is applied to radar imagery from 2011 to 2016 for the contiguous United States and achieved a precision of 0.848. Hail storms are primarily detected through the visual interpretation of radar imagery (Mrozet al., 2017). With radars providing data every two minutes, the detection of hail storms has become a big data task. As a result, scientists have turned to neural networks that employ computer vision to identify hail-bearing storms (Marzbanet al., 2001). In this study, we propose a deep Convolutional Neural Network (ConvNet) to understand the spatial features and patterns of radar echoes for detecting hailstorms.

natural hazard↗

NASA NTRS2017

Optical Flow for Flight and Wind Tunnel Background Oriented Schlieren Imaging

Background oriented Schlieren images have historically been generated by calculating the observed pixel displacement between a wind-on and wind-o image pair using normalized cross-correlation. This work uses optical flow to solve the displacement fields which generate the Schlieren images. A well established method used in the computer vision community, optical flow is the apparent motion in an image sequence due to brightness changes. The regularization method of Horn and Schunck is used to create Schlieren images using two data sets: a supersonic jet plume shock interaction from the NASA Ames Unitary Plan Wind Tunnel, and a transonic flight test of a T-38 aircraft using a naturally occurring background, performed in conjunction with NASA Ames and Armstrong Research Centers. Results are presented and contrasted with those using normalized cross-correlation. The optical flow Schlieren images are found to provided significantly more detail. We apply the method to historical data sets to demonstrate the broad applicability and limitations of the technique.

Image processing↗

NASA NTRS2015

Automated Segmentation and Classification of Coral Using Fluid Lensing from Unmanned Airborne Platforms

In recent years, there has been a growing interest among biologists in monitoring the short and long term health of the worlds coral reefs. The environmental impact of climate change poses a growing threat to these biologically diverse and fragile ecosystems, prompting scientists to use remote sensing platforms and computer vision algorithms to analyze shallow marine systems. In this study, we present a novel method for performing coral segmentation and classification from aerial data collected from small unmanned aerial vehicles (sUAV). Our method uses Fluid Lensing algorithms to remove and exploit strong optical distortions created along the air-fluid boundary to produce cm-scale resolution imagery of the ocean floor at depths up to 5 meters. A 3D model of the reef is reconstructed using structure from motion (SFM) algorithms, and the associated depth information is combined with multidimensional maximum a posteriori (MAP) estimation to separate organic from inorganic material and classify coral morphologies in the Fluid-Lensed transects. In this study, MAP estimation is performed using a set of manually classified 100 x 100 pixel training images to determine the most probable coral classification within an interrogated region of interest. Aerial footage of a coral reef was captured off the coast of American Samoa and used to test our proposed method. 90 x 20 meter transects of the Samoan coastline undergo automated classification and are manually segmented by a marine biologist for comparison, leading to success rates as high as 85. This method has broad applications for coastal remote sensing, and will provide marine biologists access to large swaths of high resolution, segmented coral imagery.

algorithms segments classifications↗

NASA NTRS2019

Planetary Rover Simulation for Lunar Exploration Missions

When planning planetary rover missions it is useful to develop intuition and skills driving in, quite literally, alien environments before incurring the cost of reaching said locales. Simulators make it possible to operate in environments that have the physical characteristics of target locations without the expense and overhead of extensive physical tests. To that end, NASA Ames and Open Robotics collaborated on a Lunar rover driving simulator based on the open source Gazebo simulation platform and leveraging ROS (Robotic Operating System) components. The simulator was integrated with research and mission software for rover driving, system monitoring, and science instrument simulation to constitute an end-to-end Lunar mission simulation capability. Although we expect our simulator to be applicable to arbitrary Lunar regions, we designed to a reference mission of prospecting in polar regions. The harsh lighting and low illumination angles at the Lunar poles combine with the unique reflectance properties of Lunar regolith to present a challenging visual environment for both human and computer perception. Our simulator placed an emphasis on high fidelity visual simulation in order to produce synthetic imagery suitable for evaluating human rover drivers with navigation tasks, as well as providing test data for computer vision software development.In this paper, we describe the software used to construct the simulated Lunar environment and the components of the driving simulation. Our synthetic terrain generation software artificially increases the resolution of Lunar digital elevation maps by fractal synthesis and inserts craters and rocks based on Lunar size-frequency distribution models. We describe the necessary enhancements to import large scale, high resolution terrains into Gazebo, as well as our approach to modeling the visual environment of the Lunar surface. An overview of the mission software system is provided, along with how ROS was used to emulate flight software components that had not been developed yet. Finally, we discuss the effect of using the high-fidelity synthetic Lunar images for visual odometry. We also characterize the wheel slip model, and find some inconsistencies in the produced wheel slip behaviour.

Allan, Mark↗

NASA NTRS2020

Arcjet Supplemental Diagnostics

This document proposes a new set of diagnostics designed to be implemented on the NASA Ames miniature Arcjet Research Chamber (mARC) for improved characterization of the flow. The diagnostics are grouped into three classes:higher cadence measurements, higher spatial resolution, and computer vision techniques for improved analysis of existing imaging. The goal is to better understand and quantify the following properties: flow statistics/uncertainty, temporal & spatial non-uniformity, flow temperature/enthalpy

Haw, Magnus A.↗

NASA NTRS2020

Deep Learning Models for Planetary Seismicity Detection

Research in planetary seismology is fundamentally constrained by a lack of data. Seismo-logical science products of future missions can typically only be informed by theoretical signal/noise characteristics of the environment or likely Earth-analogues. Although objectives can be re-assessed after some initial data-collection upon lander arrival, transfer of high-resolution data back to Earth is costly on lander power usage. Over the last several years, development of GPU computing techniques and open-source high-level APIs have led to rapid advances in deep learning within the fields of computer vision, natural language processing, and collaborative filtering. These techniques are actively being adapted in seismology for a variety of tasks, including: earthquake detection, seismic phase discrimination, and ground-motion prediction. Until the recent detection of mars quakes during the Mars InSight mission, the only other measurements of seismicity recorded outside of Earth was on the Moon during the Apollo missions between 1969 to 1977. These unique data sets have been periodically revisited using new seismological methods, including ambient noise interferometry and Hidden Markov Models. Our objective is to develop a deep learning seismic detector and use it to catalog moonquakes from the Apollo 17 Lunar Seismic Profiling Experiment (LSPE) and compare the results with those obtained by other methods. Additionally, we will assess the accuracy tradeoff between using a training set of lunar data and one composed of Earth seismicity. In this document, we present preliminary results using a prototype classifier trained on a small set of earthquakes that was able to obtain detections for LSPE moonquakes with a greater accuracy than a recent study using Hidden Markov Models.

Civilini, F.↗

NASA NTRSDate not supplied

One Giant Leap for Womankind: An Aerodynamic Study of the SLS Rocket ft. Pressure-Sensitive Paint

Combining computer vision techniques, high-speed cameras, pressure-sensitive paint and a transonic wind tunnel, NASA studies unsteady aerodynamic forces on the Space Launch System rocket with unprecedented temporal and spatial resolution. NASA’s most powerful supercomputer, Pleiades, enables parallel processing and real-time visualization to investigate buffet forces and aeroacoustic physics.

Lucy Tang↗

NASA NTRSDate not supplied

Validation of Image-Based Neural Network Controllersthrough Adaptive Stress Testing

Neural networks have become state-of-the-art for computer vision problems because of their ability to efficiently model complex functions from large amounts of data. While neural networks can be shown to perform well empirically fora variety of tasks, their performance is difficult to guarantee.Neural network verification tools have been developed that can certify robustness with respect to a given input image; however,for neural network systems used in closed-loop controllers,robustness with respect to individual images does not address multi-step properties of the neural network controller and itsenvironment. Furthermore, neural network systems interacting in the physical world and using natural images are operating in a black-box environment, making formal verification in-tractable. This work combines the adaptive stress testing (AST)framework with neural network verification tools to search for the most likely sequence of image disturbances that cause the neural network controlled system to reach a failure. Anautonomous aircraft taxi application is presented, and results show that the AST method finds failures with more likely image disturbances than baseline methods. Further analysis of AST results revealed an explainable cause of the failure, giving insight into the problematic scenarios that should be addressed.

Adaptive Stress Testing, Marabou, Deep Neural Netw↗

NASA NTRSDate not supplied

One Giant Leap for Womankind: Studying NASA’s SLS Rocket With Pressure-Sensitive Paint

Combining computer vision techniques, high-speed cameras, pressure-sensitive paint and a transonic wind tunnel, NASA studies unsteady aerodynamic forces on the Space Launch System rocket with unprecedented temporal and spatial resolution. NASA’s most powerful supercomputer, Pleiades, enables real-time visualization and parallel processing to investigate buffet forces and aeroacoustic physics.

Lucy Zhonghui Tang↗

NASA NTRSDate not supplied

Fluid for Thought: SLS Rocket Aerodynamics featuring Pressure Sensitive Paint

Combining computer vision algorithms, state-of-the-art high-speed cameras, pressure sensitive paint (PSP) and transonic wind tunnel facilities, teams at NASA and Boeing study unsteady pressure over the Space Launch System (SLS) rocket model with unprecedented temporal and spatial data resolution. Post-processing visualizes pressure fluctuations over a 3D model by converting intensity data from videos of the wind tunnel test. Additionally, data analysis creates predictions of buffet forcing and aeroacoustic activity over the rocket body with the accuracy of a CFD model! The development of the technique will revolutionize aerodynamics research and reveal uncharted insights into fluid behavior and vehicle performance, enhancing aerospace technology and design as they advance. The talk will focus on the impact of the PSP technique on the SLS design and other potential aerospace applications, as well as lay out the experimental process, from wind tunnel test to insights from the data analysis.

Lucy Zhonghui Tang↗

NASA NTRS2022

A Novel Machine Learning-Based Gap-Filling of Fine-Resolution Remotely Sensed Snow Cover Fraction Data By Combining Downscaling and Regression

Satellite-based remotely sensed observations of snow cover fraction (SCF) can have data gaps in spatially distributed coverage from sensor and orbital limitations. We mitigate these limitations in the example fine-resolution Moderate Resolution Imaging Spectroradiometer (MODIS) data by gap-filling using auxiliary 1-km datasets that either aid in downscaling from coarser-resolution (5 km) MODIS SCF wherever not fully covered by clouds, or else by themselves via regression wherever fully cloud covered. This study’s prototype predicts a 1-km version of the 500-m MOD10A1 SCF target. Due to noncollocatedness of spatial gaps even across input and auxiliary datasets, we consider a recent gap-agnostic advancement of partial convolution in computer vision for both training and predictive gap-filling. Partial convolution accommodates spatially consistent gaps across the input images, effectively implementing a two-dimensional masking. To overcome reduced usable data from noncollocated spatial gaps across inputs, we innovate a fully generalized three-dimensional masking in this partial convolution. This enables a valid output value at a pixel even if only a single valid input variable and its value exist in the neighborhood covered by the convolutional filter zone centered around that pixel. Thus, our gap-agnostic technique can use significantly more examples for training (∼67%) and prediction (∼100%), instead of only less than 10% for the previous partial convolution. We train an example simple three-layer legacy super-resolution convolutional neural network (SRCNN) to obtain downscaling and regression component performances that are better than baseline values of either climatology or MOD10C1 SCF as relevant. Our generalized partial convolution can enable multiple Earth science applications like downscaling, regression, classification, and segmentation that were hindered by data gaps.

Soni Yatheendradas↗

NASA NTRSDate not supplied

ArcjetCV: a new machine learning application for extracting time-resolved recession measurements from arc jet test videos

Arc jet Computer Vision (ArcjetCV) is a software application built to automate analysis of arc jet ground test video footage. This includes tracking material recession, sting arm motion, and the shock-material standoff distance. This provides a new capability to resolve and validate new physics associated with non-linear processes. This is an essential step to reduce testing, modeling, and validation uncertainties for heatshield material performance. ArcjetCV uses several types of machine learning (convolutional neural net: CNN, decision tree: DT, k-means unsupervised clustering: KM) to automate the video processing pipeline. These include inferring the start/stop of time segments of interest (1D CNN), measuring the time-dependent 2D recession of the material samples (2D CNN, DT), measuring the time-dependent shock standoff distance (2D CNN, DT), and post-processing cleaning of the recession data (KM). The software also provides a graphical user interface for ease of use. The results of using this tool on arc jet videos show non-linear time-dependent effects can be important for certain materials.

machine learning↗

NASA NTRSDate not supplied

Automated Tracking of Shallow Maritime Clouds on Geostationary Imagery to Extract Lifecycle Characteristics

Shallow moist convection is ubiquitous throughout the tropics and represents a key player in boundary layer processes. Satellites have provided many statistics on shallow clouds, such as size, structure, and geographical coverage, from static views of recurring cloud fields. But determining why certain cloud features appear and persist for different periods requires a time-evolving view of their behaviors. Geostationary satellites provide a unique opportunity to follow the time evolution of individual convective features, given their enhanced spatial and temporal sampling. A cloud-tracking tool was developed to identify properties of cloud lifecycle from the NASA Cloud, Aerosol, and Monsoon Processes Philippines Experiment (CAMP2EX) field campaign of 2019. The mission conducted intensive sampling of shallow cumulus in the West Pacific Ocean, in tandem with Rapid Scan imagery from the Advanced Himawari Imager (AHI) on the Japan Meteorological Agency’s (JMA) Himawari-8 satellite. Shallow cumulus was segmented according to thresholds in 0.5-km visible reflectance and with blurring techniques. Despite being limited to daytime hours, the segmentations yielded the best resolution possible for capturing cloud initiation and decay. The tracking procedure is based on a computer vision package that includes Kalman filters for motion prediction, object overlap search, and the Hungarian (or Kuhn-Munkres) matching algorithm for track designation. AHI radiances available within the tracked cloud boundaries are assembled to form individual spectral histories. The resulting catalog provides thousands of cloud histories for domains measuring only a few degrees in latitude and longitude. We present an overview of the cloud-tracking tool, strategies to identify development stages from cloud tracks, and preliminary results that document cumulus lifecycle properties from satellite. The application of AHI 0.5-km reflectance has both strengths and limitations when attempting to track lifecycles of the smallest resolvable clouds. We show that by aggregating cloud tracks from a few case studies of CAMP2EX, we can discern differences in cloud lifetime and development according to ensembles selected from areas of interest. The results demonstrate an ability to quantify lifetimes and assess rates of change in cloud characteristics that are likely controlled by the surrounding environment and meteorology.

Cloud Tracking↗

NASA NTRSDate not supplied

ArcjetCV: A New Machine Learning Application for Extracting Time-Resolved Recession Measurements From Arc Jet Test Videos

machine learning↗

NASA NTRS2022

IRIS: High-fidelity Perception Sensor Modeling for Closed-Loop Planetary Simulations

Perception plays a key role in autonomous and semi-autonomous planetary exploration vehicles. For instance, landers can use computer vision techniques for identifying safe landing locations, aerial vehicles use cameras as navigation sensors, and planetary rovers use them for localization and hazard detection. Engineering simulations of such systems requires the accurate modeling of perception and vision sensors for simulating autonomy scenarios. In addition, the modeling of sensors for landers, aerial and ground vehicles requires the ability to handle large and high-resolution terrains, the accurate modeling of illumination, hi-fidelity rendering via ray/path tracing and the inclusion of sensor characteristics. Vision sensor models strive to simulate sensor reality by using physics principles to model the interaction of light and objects. Furthermore, high frame rate performance is highly desirable for in-the-loop simulations involving vehicle dynamics and control software. In this paper we describe a new sensor modeling capability called Inter-planetary Rendering for Imaging and Sensors (IRIS) that meets these requirements for the real-time and high-fidelity simulation of vision sensors for planetary aerospace and robotics applications.

Elmquist, Asher↗

NASA NTRS2022

Velocimeter LIDAR Based Bulk Velocity Estimation for Terrain Relative Navigation Applications

A novel batch state estimation approach to estimate the translational and angular velocity states of a vehicle using a state-of-the-art velocimeter Light Detection and Ranging (LIDAR) sensor for use in Terrain and Hazard Relative Navigation (TRN/HRN) applications is presented in this paper. The velocimeter LIDAR is capable of measuring three dimensional position and line-of-sight (LOS) velocity associated with every pixel in the field-of-view (FOV). This new batch state estimation methodology is shown to provide accurate and statistically consistent velocity estimates with no a priori information. In contrast to traditional computer vision approaches for TRN, the proposed technique is not dependent on image features. This feature alleviates the need for accurate feature detection and correspondence to a predefined map making it suitable for unknown operating environments. Following a detailed development of the mathematical details associated with the batch state estimation methodology, the efficacy and utility of the proposed algorithms are evaluated through emulation robotics experiments performed at Texas A&M’s Land, Air, and Space Robotics (LASR) laboratory.

Skulsky, Eli↗