Automatic interpretation and classification of images by use of the Fourier domain
Automatic image interpretation and classification using two dimensional digital Fourier transforms
Andrews, H. C.↗
SEARCH · Engineering Papers
Results for “Image Transforms”
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.
Image processing techniques and applications to the Earth Resources Technology Satellite program
The Earth Resources Technology Satellite system is studied, with emphasis on sensors, data processing requirements, and image data compression using the Fast Fourier and Hadamard transforms. The ERTS-A system and the fundamentals of remote sensing are discussed. Three user applications (forestry, crops, and rangelands) are selected and their spectral signatures are described. It is shown that additional sensors are needed for rangeland management. An on-board information processing system is recommended to reduce the amount of data transmitted.
Polge, R. J.↗
Bi-resampled data study
The results of an experimental study conducted to determine the geometric and radiometric effects of double resampling (bi-resampling) performed on image data in the process of performing map projection transformations are reported.
Benner, R.↗
Use of satellite data in agricultural surveys
The state-of-the-art of crop surveying by satellite is reviewed with an emphasis on the signature extension problem. Registration and preprocessing procedures are discussed with refereence to: normalization of the radiometric values of each scene for scene-to-scene differences; registration techniques, implemented at the NASA Johnson Space Center, capable of 0.5 pixel root-mean-square error; and current research in this direction. Data transformation and modeling techniques applied to the Landsat MSS images and a solution for the field-to-field variations of the greenness and brightness temporal trajectories are included. Finally, a review of the mixture decomposition method of labeling and estimating the areal proportions is given.
Hall, F. G.↗
Shiftable multiscale transforms
A type of translational invariance, referred to as shiftability, is defined for wavelet transforms. The property of shiftability is first discussed with respect to individual parameters: spatial position, orientation, and scale. The discussion then focuses on transformations that are simultaneously shiftable with respect to subsets of these parameters. It is shown that the critical sampling condition on the wavelet transforms must be relaxed to achieve shiftability. Two example transforms are implemented and applied to several signal and image processing problems.
Simoncelli, Eero P.↗
Mesospheric and lower thermospheric winds, temperatures densities, and volume emission rates
This memo has been written to report on the progress made under grant NAG1-1315 and to request a continuation of funding for this work. Our proposal involved a plan to utilize an existing ground-based chain of optical and radar facilities to assemble a comprehensive, long-term, multi-station base of upper-mesospheric and lower thermospheric measurements of neutral winds, temperatures, and volume emission rates that can be used to make comparisons with data from the HRDI and WINDII instruments that are flying aboard the UARS spacecraft. The ground-based, optical data were to be obtained on a routine basis from five geographically separated observatories at: Thule, Greenland; Sondrestrom, Greenland; Watson Lake, Yukon (replacing Calgary, Alberta); Ann Arbor, Michigan; and Maynooth, Eire. Several different optical instruments are present at these sites: the total array of instruments is comprised of five Fabry-Perot interferometers, two half meter Ebert-Fastie spectrometers, one all-sky CCD imager, and a near infra-red Michelson Fourier transform spectrometer. In addition to these optical instruments, data were to be obtained from the incoherent scatter radar at Sondrestrom, Greenland. These radar measurements are comprised of neutral winds, temperatures, and densities from altitudes between approximately 70 - 120 km. The optical measurements are obtained locally from specific altitudes depending on the emission line studied. For example, red line optical data come from about 220 km. In this report we summarize the progress made in obtaining these data and relate it to the specific tasks outlined in the original grant application. These tasks are discussed in the next section. Progress towards their completion is discussed in section three, while future plans and summary are described in section four.
Killeen, Timothy L.↗
The Science Performance of JWST as Characterized in Commissioning
This paper characterizes the actual science performance of the James Webb Space Telescope (JWST), as determined from the six month commissioning period. We summarize the performance of the spacecraft, telescope, science instruments, and ground system, with an emphasis on differences from pre-launch expectations. Commissioning has made clear that JWST is fully capable of achieving the discoveries for which it was built. Moreover, almost across the board, the science performance of JWST is better than expected; in most cases, JWST will go deeper faster than expected. The telescope and instrument suite have demonstrated the sensitivity, stability, image quality, and spectral range that are necessary to transform our understanding of the cosmos through observations spanning from near-earth asteroids to the most distant galaxies.
Infrared astronomy↗
A Method for Obtaining Surface Flow Vectors and Its Implementation in Interferometric Skin Friction Measurement
A new method was developed to extract surface flow vectors from an oilflow visualization image that has oil streaklines. The method is analogous to the PIV processing where the image is divided into interrogation windows. A representative flow direction is obtained for each interrogation window using image processing with a line-detection algorithm. Repeating the process for the entire image, one can find the surface vector field. The Hough transformation, Radon transformation and Machine Learning were used as line-detection algorithms. The obtained vector field is then postprocessed to filter spurious vectors and to apply smoothing. The method was tested on 2D and 3D models with different flow complexities. The method was able to predict surface flow vectors for all cases tested. The predicted surface flow vectors were used in obtaining the surface skin friction. The successful implementation of the surface flow vectors enables interferometric skin friction measurements on surfaces beneath 3D complex flows. In addition, the surface flow vectors can be superimposed on the oilflow visualization images to better explain the surface flow topology.
Surface flow vectors↗
A Method for Obtaining Surface Flow Vectors and Its Implementation in Interferometric Skin Friction Measurement
A new method was developed to extract surface flow vectors from an oilflow visualization image that has oil streaklines. The method is analogous to the PIV processing where the image is divided into interrogation windows. A representative flow direction is obtained for each interrogation window using image processing with a line-detection algorithm. Repeating the process for the entire image, one can find the surface vector field. The Hough transformation, Radon transformation and Machine Learning were used as line-detection algorithms. The obtained vector field is then postprocessed to filter spurious vectors and to apply smoothing. The method was tested on 2D and 3D models with different flow complexities. The method was able to predict surface flow vectors for all cases tested. The predicted surface flow vectors were used in obtaining the surface skin friction. The successful implementation of the surface flow vectors enables interferometric skin friction measurements on surfaces beneath 3D complex flows. In addition, the surface flow vectors can be superimposed on the oilflow visualization images to better explain the surface flow topology.
Surface flow vectors↗
Image Tiling for Profiling Large Objects
Three dimensional surface measurements of large objects arc required in a variety of industrial processes. The nature of these measurements is changing as optical instruments arc beginning to replace conventional contact probes scanned over the objects. A common characteristic of the optical surface profilers is the trade off between measurement accuracy and field of view. In order to measure a large object with high accuracy, multiple views arc required. An accurate transformation between the different views is needed to bring about their registration. In this paper, we demonstrate how the transformation parameters can be obtained precisely by choosing control points which lie in the overlapping regions of the images. A good starting point for the transformation parameters is obtained by having a knowledge of the scanner position. The selection of the control points arc independent of the object geometry. By successively recording multiple views and obtaining transformation with respect to a single coordinate system, a complete physical model of an object can be obtained. Since all data arc in the same coordinate system, it can thus be used for building automatic models for free form surfaces.
Venkataraman, Ajit↗
Predicting Nugget Size of Resistance Spot Welds Using Infrared Thermal Videos With Image Segmentation and Convolutional Neural Network
Resistance spot welding (RSW) is a widely adopted joining technique in automotive industry. Recent advancement in sensing technology makes it possible to collect thermal videos of the weld nugget during RSW using an infrared (IR) camera. The effective and timely analysis of such thermal videos has the potential of enabling in situ nondestructive evaluation (NDE) of the weld nugget by predicting nugget thickness and diameter. Deep learning (DL) has demonstrated to be effective in analyzing imaging data in many applications. However, the thermal videos in RSW present unique data-level challenges that compromise the effectiveness of most pre-trained DL models. We propose a novel image segmentation method for handling the RSW thermal videos to improve the prediction performance of DL models in RSW. The proposed method transforms raw thermal videos into spatial-temporal instances in four steps: video-wise normalization, removal of uninformative images, watershed segmentation, and spatial-temporal instance construction. The extracted spatial-temporal instances serve as the input data for training a DL-based NDE model. The proposed method is able to extract high-quality data with spatial-temporal correlations in the thermal videos, while being robust to the impact of unknown surface emissivity. Overall, our case studies demonstrate that the proposed method achieves better prediction of nugget thickness and diameter than predicting without the transformation.
42 ENGINEERING↗
Hybrid Quantum Vision Transformers for Event Classification in High Energy Physics
Models based on vision transformer architectures are considered state-of-the-art when it comes to image classification tasks. However, they require extensive computational resources both for training and deployment. The problem is exacerbated as the amount and complexity of the data increases. Quantum-based vision transformer models could potentially alleviate this issue by reducing the training and operating time while maintaining the same predictive power. Although current quantum computers are not yet able to perform high-dimensional tasks, they do offer one of the most efficient solutions for the future. In this work, we construct several variations of a quantum hybrid vision transformer for a classification problem in high-energy physics (distinguishing photons and electrons in the electromagnetic calorimeter). We test them against classical vision transformer architectures. Our findings indicate that the hybrid models can achieve comparable performance to their classical analogs with a similar number of parameters.
Unlu, Eyup B. (ORCID:0000000266836463)↗
Computer program for fast Karhunen Loeve transform algorithm
The fast KL transform algorithm was applied for data compression of a set of four ERTS multispectral images and its performance was compared with other techniques previously studied on the same image data. The performance criteria used here are mean square error and signal to noise ratio. The results obtained show a superior performance of the fast KL transform coding algorithm on the data set used with respect to the above stated perfomance criteria. A summary of the results is given in Chapter I and details of comparisons and discussion on conclusions are given in Chapter IV.
Jain, A. K.↗
Calibration and characterization of the line-VISAR diagnostic at the HED-HIBEF instrument at the European XFEL
In dynamic-compression experiments, the line-imaging Velocity Interferometer System for Any Reflector (VISAR) is a well-established diagnostic used to probe the velocity history, including wave profiles derived from dynamically compressed interfaces and wavefronts, depending on material optical properties. Knowledge of the velocity history allows for the determination of the pressure achieved during compression. Such a VISAR analysis is often based on Fourier transform techniques and assumes that the recorded interferograms are free from image distortions. In this paper, we describe the VISAR diagnostic installed at the HED-HIBEF instrument located at the European XFEL along with its calibration and characterization. It comprises a two-color (532, 1064 nm), three-arm (with three velocity sensitivities) line imaging system. We provide a procedure to correct VISAR images for geometric distortions and evaluate the performance of the system using Fourier analysis. We finally discuss the spatial and temporal calibrations of the diagnostic. As an example, we compare the pressure extracted from the VISAR analysis of shock-compressed polyimide and silicon.
47 OTHER INSTRUMENTATION↗
Visual optimization of DCT quantization matrices for individual images
Many image compression standards (JPEG, MPEG, H.261) are based on the Discrete Cosine Transform (DCT). However, these standards do not specify the actual DCT quantization matrix. We have previously provided mathematical formulae to compute a perceptually lossless quantization matrix. Here I show how to compute a matrix that is optimized for a particular image. The method treats each DCT coefficient as an approximation to the local response of a visual 'channel'. For a given quantization matrix, the DCT quantization errors are adjusted by contrast sensitivity, light adaptation, and contrast masking, and are pooled non-linearly over the blocks of the image. This yields an 8x8 'perceptual error matrix'. A second non-linear pooling over the perceptual error matrix yields total perceptual error. With this model we may estimate the quantization matrix for a particular image that yields minimum bit rate for a given total perceptual error, or minimum perceptual error for a given bit rate. Custom matrices for a number of images show clear improvement over image-independent matrices. Custom matrices are compatible with the JPEG standard, which requires transmission of the quantization matrix.
Watson, Andrew B.↗
Hybrid model to evaluate the frequency-dependent leakage inductance of partially-filled transformers
The leakage inductance of a transformer designed for a power electronic converter can drop significantly as the switching frequency is increased due to skin and proximity effects. Although the magnetic image method-based double-2-D model can predict the low-frequency leakage inductance of a partially-filled transformer with sufficient accuracy, it is inherently a frequency-independent model. While Dowell’s 1-D model uses frequency-dependent relations to account for both skin and proximity effects, its accuracy is severely affected by the assumed winding geometry. In this paper, a semianalytical hybrid model is proposed that uses superposition to combine a modified Dowell’s model with the double-2-D model to predict the true leakage inductance of partially-filled transformers at any given frequency. All three conductor types—round, foil, and litz wire—are modeled and analyzed. The quasi-2-D model is further investigated on a variable inductance transformer (VIT) whose winding geometry can be modified mechanically to vary its leakage inductance. With less than 5 % error throughout, the semi-analytically evaluated leakage inductances are in excellent agreement with the finite element method (FEM) simulated and experimentally measured leakage inductances.
24 POWER TRANSMISSION AND DISTRIBUTION↗
Self calibrating autoTRAC
The work reported here demonstrates how to automatically compute the position and attitude of a targeting reflective alignment concept (TRAC) camera relative to the robot end effector. In the robotics literature this is known as the sensor registration problem. The registration problem is important to solve if TRAC images need to be related to robot position. Previously, when TRAC operated on the end of a robot arm, the camera had to be precisely located at the correct orientation and position. If this location is in error, then the robot may not be able to grapple an object even though the TRAC sensor indicates it should. In addition, if the camera is significantly far from the alignment it is expected to be at, TRAC may give incorrect feedback for the control of the robot. A simple example is if the robot operator thinks the camera is right side up but the camera is actually upside down, the camera feedback will tell the operator to move in an incorrect direction. The automatic calibration algorithm requires the operator to translate and rotate the robot arbitrary amounts along (about) two coordinate directions. After the motion, the algorithm determines the transformation matrix from the robot end effector to the camera image plane. This report discusses the TRAC sensor registration problem.
Everett, Louis J.↗
Object detection in natural backgrounds predicted by discrimination performance and models
Many models of visual performance predict image discriminability, the visibility of the difference between a pair of images. We compared the ability of three image discrimination models to predict the detectability of objects embedded in natural backgrounds. The three models were: a multiple channel Cortex transform model with within-channel masking; a single channel contrast sensitivity filter model; and a digital image difference metric. Each model used a Minkowski distance metric (generalized vector magnitude) to summate absolute differences between the background and object plus background images. For each model, this summation was implemented with three different exponents: 2, 4 and infinity. In addition, each combination of model and summation exponent was implemented with and without a simple contrast gain factor. The model outputs were compared to measures of object detectability obtained from 19 observers. Among the models without the contrast gain factor, the multiple channel model with a summation exponent of 4 performed best, predicting the pattern of observer d's with an RMS error of 2.3 dB. The contrast gain factor improved the predictions of all three models for all three exponents. With the factor, the best exponent was 4 for all three models, and their prediction errors were near 1 dB. These results demonstrate that image discrimination models can predict the relative detectability of objects in natural scenes.
NASA Center ARC↗