Search NASA⌕ Search

SEARCH · Search NASA

Results for “map segmentation”

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.

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

At least 37 records · Page 2

A Semiautomated Multilayer Picking Algorithm for Ice-sheet Radar Echograms Applied to Ground-Based Near-Surface Data

Snow accumulation over an ice sheet is the sole mass input, making it a primary measurement for understanding the past, present, and future mass balance. Near-surface frequency-modulated continuous-wave (FMCW) radars image isochronous firn layers recording accumulation histories. The Semiautomated Multilayer Picking Algorithm (SAMPA) was designed and developed to trace annual accumulation layers in polar firn from both airborne and ground-based radars. The SAMPA algorithm is based on the Radon transform (RT) computed by blocks and angular orientations over a radar echogram. For each echogram's block, the RT maps firn segmented-layer features into peaks, which are picked using amplitude and width threshold parameters of peaks. A backward RT is then computed for each corresponding block, mapping the peaks back into picked segmented-layers. The segmented layers are then connected and smoothed to achieve a final layer pick across the echogram. Once input parameters are trained, SAMPA operates autonomously and can process hundreds of kilometers of radar data picking more than 40 layers. SAMPA final pick results and layer numbering still require a cursory manual adjustment to correct noncontinuous picks, which are likely not annual, and to correct for inconsistency in layer numbering. Despite the manual effort to train and check SAMPA results, it is an efficient tool for picking multiple accumulation layers in polar firn, reducing time over manual digitizing efforts. The trackability of good detected layers is greater than 90%.

Antarctic Ice-Sheet↗

Semantic Segmentation of High-Resolution Satellite Imagery using Generative Adversarial Networks with Progressive Growing

With increase in urbanization and Earth Sciences research into urban areas, the need to quickly and accurately segment urban rooftop maps has never been greater. Cur-rent machine learning techniques struggle to produce high accuracy maps in dense urban zones where there is high image noise and foot print overlap. In this paper, we evaluate a training methodology for pixel-wise segmentation for high resolution satellite imagery using progressive growing of generative adversarial networks as a solution. We apply our model to segmenting building rooftops and compare these results to conventional methods for rooftop segmentation. We evaluate our approach using the SpaceNet version 2 and xView datasets. Our experiments show that for SpaceNet, progressive Generative Adversarial Network (GAN) training achieved a test accuracy of 93% compared to 89% for traditional GAN training and 87% for U-Net architecture, while for xView, we achieved 71% accuracy using progressive GAN training compared to 69% through traditional GAN training and 65% using U-Net.

Semantic↗

Technology Advancement of the Visible Nulling Coronagraph

The critical high contrast imaging technology for the Extrasolar Planetary Imaging Coronagraph (EPIC) mission concept is the visible nulling coronagraph (VNC). EPIC would be capable of imaging jovian planets, dust/debris disks, and potentially super-Earths and contribute to answering how bright the debris disks are for candidate stars. The contrast requirement for EPIC is 10(exp 9) contrast at 125 milli-arseconds inner working angle. To advance the VNC technology NASA/Goddard Space Flight Center, in collaboration with Lockheed-Martin, previously developed a vacuum VNC testbed, and achieved narrowband and broadband suppression of the core of the Airy disk. Recently our group was awarded a NASA Technology Development for Exoplanet Missions to achieve two milestones: (i) 10(exp 8) contrast in narrowband light, and, (ii) 10(ecp 9) contrast in broader band light; one milestone per year, and both at 2 Lambda/D inner working angle. These will be achieved with our 2nd generation testbed known as the visible nulling testbed (VNT). It contains a MEMS based hex-packed segmented deformable mirror known as the multiple mirror array (MMA) and coherent fiber bundle, i.e. a spatial filter array (SFA). The MMA is in one interferometric arm and works to set the wavefront differences between the arms to zero. Each of the MMA segments is optically mapped to a single mode fiber of the SFA, and the SFA passively cleans the sub-aperture wavefront error leaving only piston, tip and tilt error to be controlled. The piston degree of freedom on each segment is used to correct the wavefront errors, while the tip/tilt is used to simultaneously correct the amplitude errors. Thus the VNT controls both amplitude and wavefront errors with a single MMA in closed-loop in a vacuum tank at approx.20 Hz. Herein we will discuss our ongoing progress with the VNT.

Lyon, Richard G.↗

Australia ground data collection 1981-82 crop year, volume 1

Under AgRISTARS management, ground data were collected at 20 agricultural sites within Australia during the crop year 1981-82. The data collection activity is summarized. Specifically, the following information is provided: discussion of data procedures, methods, and products; crop production results; photographs of the Australia agriculture scene, map sheets of segments, LANDSAT full frames, and aerial photographs of data collection areas; and summarizations of district agronomist reports.

Quinones, C. R.↗

Detecting and avoiding collisions between two robot arms in a common workspace

A novel approach to collision detection is presented and methods of avoiding collisions are discussed. Collisions are assumed to be between the wrists of two robots (which correspond to the upper three links of PUMA manipulators). A sphere model for the wrist is employed because it is rotationally invariant and computationally efficient. The collision detection algorithm involves obtaining the potential collision path segments along the straight line trajectories without considering the motion characteristics, and mapping the potential collision segment information into the time domain to obtain the space-time collisions.

Basta, Robert A.↗

The effect of scene content on speed, time, and distance perception

Helicopter flights performed at low-levels place high demands on pilots; they must simultaneously control the vehicle, avoid obstacles, and navigate. Therefore, pilots must correlate cues viewed in the external scene with information on map in order to maintain their geographical orientation. This is a particularly difficult task when helicopter pilots fly through visually unfamiliar terrain without highly detailed maps. As a result, pilots must often use estimates of elapsed time, distance traveled, and/or average speed in order to maintain a flight path indicated on a map during flight segments when these cues are absent. Therefore, the current study is concerned with the perception of speed, time, and distance, which we assume underlies the ability to orient oneself during these types of flight segments.

Awe, Cynthia A.↗

Correlations among the Optical Properties of Cirrus-Cloud Particles: Microphysical Interpretation

Cirrus measurements obtained with a ground-based polarization Raman lidar at 67.9 deg N in January 1997 reveal a strong positive correlation between the particle optical properties, specifically depolarization ratio delta(sub par) and extinction- to-backscatter (lidar) ratio S, for delta(sub par) less than approximately 40%, and an anti-correlation for delta(sub par) greater than approximately 40%. Over the length of the measurements the particle properties vary systematically. Initially, delta (sub par) approximately equals 60% and S approximately equals 10sr are observed. Then, with decreasing delta(sub par), S first increases to approximately 27sr (delta(sub par) approximately equals 40%) before decreasing to values around 10sr again (delta(sub par) approximately equals 20%). The analysis of lidar humidity and radiosonde temperature data shows that the measured optical properties stem from scattering by dry solid ice particles, while scattering by supercooled droplets, or by wetted or subliming ice particles can be excluded. For the microphysical interpretation of the lidar measurements, ray-tracing computations of particle scattering properties have been used. The comparison with the theoretical data suggests that the observed cirrus data can be interpreted in terms of size, shape, and, under the assumption that the lidar measurements of consecutive cloud segments can be mapped on the temporal development of a single cloud parcel moving along its trajectory, growth of the cirrus particles: Near the cloud top in the early stage of cirrus development, light scattering by nearly isometric particles that have the optical characteristics of hexagonal columns (short, column-like particles) is dominant. Over time the ice particles grow, and as the cloud base height extends to lower altitudes characterized by warmer temperatures they become morphologically diverse. For large S and depolarization values of approximately 40%, the scattering contributions of column- and plate-like particles are roughly the same. In the lower ranges of the cirrus clouds, light scattering is predominantly by plate-like ice particles. This interpretation assumes random orientation of the cirrus particles. Simulations with a simple model suggest, however, that the positive correlation between S and delta(sub par) which is observed for depolarization ratios less than 40% mainly at low cloud altitudes, can be alternatively explained by horizontal alignment of a fraction of the cirrus particle population.

Reichardt, J.↗

Metric Learning to Enhance Hyperspectral Image Segmentation

Unsupervised hyperspectral image segmentation can reveal spatial trends that show the physical structure of the scene to an analyst. They highlight borders and reveal areas of homogeneity and change. Segmentations are independently helpful for object recognition, and assist with automated production of symbolic maps. Additionally, a good segmentation can dramatically reduce the number of effective spectra in an image, enabling analyses that would otherwise be computationally prohibitive. Specifically, using an over-segmentation of the image instead of individual pixels can reduce noise and potentially improve the results of statistical post-analysis. In this innovation, a metric learning approach is presented to improve the performance of unsupervised hyperspectral image segmentation. The prototype demonstrations attempt a superpixel segmentation in which the image is conservatively over-segmented; that is, the single surface features may be split into multiple segments, but each individual segment, or superpixel, is ensured to have homogenous mineralogy.

Thompson, David R.↗

Wheat cultivation: Identification and estimation of areas using LANDSAT data

The feasibility of using automatically processed multispectral data obtained from LANDSAT to identify wheat and estimate the areas planted with this grain was investigated. Three 20 km by 40 km segments in a wheat growing region of Rio Grande do Sul were aerially photographed using type 2443 Aerochrome film. Three maps corresponding to each segment were obtained from the analysis of the photographs which identified wheat, barley, fallow land, prepared soil, forests, and reforested land. Using basic information about the fields and maps made from the photographed areas, an automatic classification of wheat was made using MSS data from two different periods: July to September and July to October 1979. Results show that orbital data is not only useful in characterizing the growth of wheat, but also provides information of the intensity and extent of adverse climate which affects cultivation. The temporal and spatial characteristics of LANDSAR data are also demonstrated.

Dejesusparada, N.↗

Geologic and mineral and water resources investigations in western Colorado, using Skylab EREP data

The author has identified the following significant results. Discovery of three major north-trending, throughgoing faults in the Front Range, previously mapped only as isolated segments, demonstrates the utility of space photography and may lead to reinterpretation of the Front Range tectonic style. Faulting and alteration appear to be the most useful indicators of mineralization in central Colorado. These phenomena appear on Skylab photography as tonal lineaments and color anomalies. Twenty-three lineaments have been mapped in the San Juan Mountains, the longest of which is 156 km long. Twelve lineaments intersect or are tangent to calderas. Intrusive domes are aligned along lineaments, but calderas appear to occur at the intersections of major lineaments. Lineaments can be recognized on some EREP passes but not on other passes over the same area. The difference is attributed to solar elevation effects. Bedding attitudes can be photogeologically estimated down to surprisingly low dips, on the order of + or - 1-2 deg, and attitudes can be subdivided easily into quantitative groups. The primary application of Skylab photography to geologic mapping in montane areas is clearly limited to regional mapping at scales smaller than 1:24,000.

Lee, K.↗

Persistence Mapping Using EUV Solar Imager Data

We describe a simple image processing technique that is useful for the visualization and depiction of gradually evolving or intermittent structures in solar physics extreme-ultraviolet imagery. The technique is an application of image segmentation, which we call "Persistence Mapping," to isolate extreme values in a data set, and is particularly useful for the problem of capturing phenomena that are evolving in both space and time. While integration or "time-lapse" imaging uses the full sample (of size N ), Persistence Mapping rejects (N - 1)/N of the data set and identifies the most relevant 1/N values using the following rule: if a pixel reaches an extreme value, it retains that value until that value is exceeded. The simplest examples isolate minima and maxima, but any quantile or statistic can be used. This paper demonstrates how the technique has been used to extract the dynamics in long-term evolution of comet tails, erupting material, and EUV dimming regions.

Thompson, B. J.↗

Bayesian Deep Learning for Segmentation for Autonomous Safe Planetary Landing

Hazard detection is critical for enabling autonomous landing on planetary surfaces. Current state-of-the-art methods leverage traditional computer vision approaches to automate the identification of safe terrain from input digital elevation models (DEMs). However, performance for these methods can degrade for input DEMs with increased sensor noise. In the last decade, deep learning techniques have been developed for various applications. Nevertheless, their applicability to safety-critical space missions has often been limited due to concerns regarding their outputs’ reliability. In response to these limitations, this paper proposes an application of the Bayesian deep learning segmentation method for hazard detection. The developed approach enables reliable, safe landing site detection by i) generating simultaneously a safety prediction map and its uncertainty map via Bayesian deep learning and semantic segmentation, and ii) using the uncertainty map to filter out the uncertain pixels in the prediction map so that the safe site identification is performed only based on the certain pixels (i.e., pixels for which the model is certain about its safety prediction). Experiments are presented with simulated data based on a Mars HiRISE digital terrain model by varying uncertainty threshold and noise levels to demonstrate the performance of the proposed approach.

Kento Tomita↗

Generating Ground Reference Data for a Global Impervious Surface Survey

We are engaged in a project to produce a 30m impervious cover data set of the entire Earth for the years 2000 and 2010 based on the Landsat Global Land Survey (GLS) data set. The GLS data from Landsat provide an unprecedented opportunity to map global urbanization at this resolution for the first time, with unprecedented detail and accuracy. Moreover, the spatial resolution of Landsat is absolutely essential to accurately resolve urban targets such as buildings, roads and parking lots. Finally, with GLS data available for the 1975, 1990, 2000, and 2005 time periods, and soon for the 2010 period, the land cover/use changes due to urbanization can now be quantified at this spatial scale as well. Our approach works across spatial scales using very high spatial resolution commercial satellite data to both produce and evaluate continental scale products at the 30m spatial resolution of Landsat data. We are developing continental scale training data at 1m or so resolution and aggregating these to 30m for training a regression tree algorithm. Because the quality of the input training data are critical, we have developed an interactive software tool, called HSegLearn, to facilitate the photo-interpretation of high resolution imagery data, such as Quickbird or Ikonos data, into an impervious versus non-impervious map. Previous work has shown that photo-interpretation of high resolution data at 1 meter resolution will generate an accurate 30m resolution ground reference when coarsened to that resolution. Since this process can be very time consuming when using standard clustering classification algorithms, we are looking at image segmentation as a potential avenue to not only improve the training process but also provide a semi-automated approach for generating the ground reference data. HSegLearn takes as its input a hierarchical set of image segmentations produced by the HSeg image segmentation program [1, 2]. HSegLearn lets an analyst specify pixel locations as being either positive or negative examples, and displays a classification of the study area based on these examples. For our study, the positive examples are examples of impervious surfaces and negative examples are examples of non-impervious surfaces. HSegLearn searches the hierarchical segmentation from HSeg for the coarsest level of segmentation at which selected positive example locations do not conflict with negative example locations and labels the image accordingly. The negative example regions are always defined at the finest level of segmentation detail. The resulting classification map can be then further edited at a region object level using the previously developed HSegViewer tool [3]. After providing an overview of the HSeg image segmentation program, we provide a detailed description of the HSegLearn software tool. We then give examples of using HSegLearn to generate ground reference data and conclude with comments on the effectiveness of the HSegLearn tool.

Tilton, James C.↗

A summary of image segmentation techniques

Machine vision systems are often considered to be composed of two subsystems: low-level vision and high-level vision. Low level vision consists primarily of image processing operations performed on the input image to produce another image with more favorable characteristics. These operations may yield images with reduced noise or cause certain features of the image to be emphasized (such as edges). High-level vision includes object recognition and, at the highest level, scene interpretation. The bridge between these two subsystems is the segmentation system. Through segmentation, the enhanced input image is mapped into a description involving regions with common features which can be used by the higher level vision tasks. There is no theory on image segmentation. Instead, image segmentation techniques are basically ad hoc and differ mostly in the way they emphasize one or more of the desired properties of an ideal segmenter and in the way they balance and compromise one desired property against another. These techniques can be categorized in a number of different groups including local vs. global, parallel vs. sequential, contextual vs. noncontextual, interactive vs. automatic. In this paper, we categorize the schemes into three main groups: pixel-based, edge-based, and region-based. Pixel-based segmentation schemes classify pixels based solely on their gray levels. Edge-based schemes first detect local discontinuities (edges) and then use that information to separate the image into regions. Finally, region-based schemes start with a seed pixel (or group of pixels) and then grow or split the seed until the original image is composed of only homogeneous regions. Because there are a number of survey papers available, we will not discuss all segmentation schemes. Rather than a survey, we take the approach of a detailed overview. We focus only on the more common approaches in order to give the reader a flavor for the variety of techniques available yet present enough details to facilitate implementation and experimentation.

Spirkovska, Lilly↗

On extracting brightness temperature maps from scanning radiometer data

The extraction of brightness temperature maps from scanning radiometer data is described as a typical linear inverse problem. Spatial quantization and parameter estimation is described and is suggested as an advantageous approach to a solution. Since this approach takes into explicit account the multivariate nature of the problem, it permits an accurate determination of the most detailed resolution extractable from the data as well as explicitly defining the possible compromises between accuracy and resolution. To illustrate the usefulness of the method described for algorithm design and accuracy prediction, it was applied to the problem of providing brightness temperature maps during the NOSS flight segment. The most detained possible resolution was determined and a curve which displays the possible compromises between accuracy and resolution was provided.

Argentiero, P.↗

As-built design specification for MISMAP

The MISMAP program, which is part of the CLASFYT package, is described. The program is designed to compare classification values with ground truth values for a segment and produce a comparison map and summary table.

Brown, P. M.↗

Robot path planning with distance-safety criterion

A method for determining an optimal path with a weighted distance-safety criterion is developed. The goal is to strike a compromise between the shortest path and the centerline path, which is safer. The method is composed of three parts: (i) construction of a region map by dividing the workspace, (ii) interregion optimization to determine the entry and departure points of the path in each region, and (iii) intraregion optimization for determining the (optimal) path segment within each region. The region map is generated by using an approximate Voronoi diagram, and region optimization is achieved using variational dynamic programming. Although developed for 2-D problems, the method can be easily extended to a class of 3-D problems. Numerical examples are presented to demonstrate the method.

Suh, Suk-Hwan↗