An official website of the United States government
In tomography, three-dimensional images of a medium are reconstructed from a set of two-dimensional projections. Each projection is the result of a measurement made by the scanner via radiating some form of energy and collecting the scattered field after interacting with the medium. The information content of these measurements is not equal, and one projection can be more informative than others. By choosing the most informative measurement at every step of scanning, an optimal tomography system can maximize the speed of data acquisition and temporal resolution of acquired images, reducing the operation cost and exposure to possible harmful radiations. The aim of this paper is to introduce mathematical algorithms that can be used to design measurements with optimal information content when imaging static or dynamically evolving objects.
more »
« less
Illumination pattern optimization in tomography based on the Kalman estimation filter and optimal experiment design
Tomography is widely used in medical imaging or industrial non-destructive testing applications. One costly and time consuming operation in any form of tomography is the process of data acquisition where a large number of measurements are made and collected data is used for image reconstruction. Data acquisition can slow down tomography to the point that the scanner cannot catch up with the speed of changes in the medium under test. By optimizing the information content of each measurement, we can reduce the number of measurements needed to achieve the target precision. Development of algorithms to optimize the information content of tomography measurements is the main goal of this article. Here, the dynamics of the medium and tomography measurements are formulated in the form of a Kalman estimation filter. A mathematical algorithm is developed to compute the optimal measurement matrix which minimizes the uncertainty left in the estimation of the distribution the tomography scanner is reconstructing. Results, as presented in the paper, show noticeable improvement is the quality of generated images when the medium is scanned by optimal measurements instead of traditional raster or random scanning protocols.
more »
« less
- Award ID(s):
- 2154267
- PAR ID:
- 10502681
- Publisher / Repository:
- Optical Society of America
- Date Published:
- Journal Name:
- Optics Express
- Volume:
- 32
- Issue:
- 10
- ISSN:
- 1094-4087; OPEXFF
- Format(s):
- Medium: X Size: Article No. 17345
- Size(s):
- Article No. 17345
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Zhang, Qichun (Ed.)We develop a general framework for state estimation in systems modeled with noise-polluted continuous time dynamics and discrete time noisy measurements. Our approach is based on maximum likelihood estimation and employs the calculus of variations to derive optimality conditions for continuous time functions. We make no prior assumptions on the form of the mapping from measurements to state-estimate or on the distributions of the noise terms, making the framework more general than Kalman filtering/smoothing where this mapping is assumed to be linear and the noises Gaussian. The optimal solution that arises is interpreted as a continuous time spline, the structure and temporal dependency of which is determined by the system dynamics and the distributions of the process and measurement noise. Similar to Kalman smoothing, the optimal spline yields increased data accuracy at instants when measurements are taken, in addition to providing continuous time estimates outside the measurement instances. We demonstrate the utility and generality of our approach via illustrative examples that render both linear and nonlinear data filters depending on the particular system. Application of the proposed approach to a Monte Carlo simulation exhibits significant performance improvement in comparison to a common existing method.more » « less
-
Classical shadows (CS) offer a resource-efficient means to estimate quantum observables, circumventing the need for exhaustive state tomography. Here, we clarify and explore the connection between CS techniques and least squares (LS) and regularized least squares (RLS) methods commonly used in machine learning and data analysis. By formal identification of LS and RLS ``shadows'' completely analogous to those in CS---namely, point estimators calculated from the empirical frequencies of single measurements---we show that both RLS and CS can be viewed as regularizers for the underdetermined regime, replacing the pseudoinverse with invertible alternatives. Through numerical simulations, we evaluate RLS and CS from three distinct angles: the tradeoff in bias and variance, mismatch between the expected and actual measurement distributions, and the interplay between the number of measurements and number of shots per measurement.Compared to CS, RLS attains lower variance at the expense of bias, is robust to distribution mismatch, and is more sensitive to the number of shots for a fixed number of state copies---differences that can be understood from the distinct approaches taken to regularization. Conceptually, our integration of LS, RLS, and CS under a unifying ``shadow'' umbrella aids in advancing the overall picture of CS techniques, while practically our results highlight the tradeoffs intrinsic to these measurement approaches, illuminating the circumstances under which either RLS or CS would be preferred, such as unverified randomness for the former or unbiased estimation for the latter.more » « less
-
Claypan soils cover approximately 10 million acres across several states in the central United States. The soils are characterized by a highly impermeable clay layer within the profile that impedes water flow and root growth. While some claypan soils can be productive, they must be carefully managed to avoid reductions to crop productivity due to root restrictions, water, and nutrient limitations. Clay soils are usually resistant to erosion but may exacerbate erosion of the silt-loam topsoil. Soil production potential is the capacity of soil to produce at a given level (yield per acre). The productive capacity is tied to soil characteristics, which can be highly variable within a field. In this project, we have used imagery analysis to study the aerial images and terrain of fields during different productive times of the year to identify where soil samples should be collected for more discrete analysis. Soil samples provide valuable information; however, the amount of data obtained from a relatively small area within a field does not provide sufficient information to delineate the subsurface characteristics. To address the limitations of sampling, we have also employed the use of yield maps collected from commercial yield monitors on production-scale combines and surface electrical conductivity measurements (Sassenrath and Kulesza, 2017). Soil conductivity is a measurement of how well a representative volume of soil conducts electricity. Soil conductivity is a function of the soil clay content, moisture content, and other measurable soil properties (Kitchen et al., 2003); as such, it has become a valuable tool for mapping in-field variability. The main advantage of a soil conductivity measurement is that the entire surface of a field can be imaged. The disadvantage of a soil conductivity measurement is that data are only collected near the surface (10 – 30 inches) and the measurements are relative measurements. This means that the conductivity mappers can identify changes in soil properties, but they cannot directly tell researchers what caused these changes. Electrical resistivity tomography (ERT) is a popular near-surface geophysical measurement for geophysical and engineering applications. The term “near-surface” generally means down to around 30 feet in the subsurface. Electrical resistivity is the reciprocal measurement of electrical conductivity; therefore, both systems measure differences in the same soil properties. ERT measurements are different than surface electrical conductivity measurements because ERT collects a “slice” of data into the subsurface, as opposed to only changes at the surface area. Relative measurements, similar to those collected in an electrical conductivity survey, are collected; however, in ERT studies the data are mathematically inverted to yield the true electrical resistivity of the soil with depth. This allows an interpretation of the changing soil properties with depth to reduce the required amount of sampling. A disadvantage of an ERT survey is that the data acquisition is stationary so mapping an entire field is not feasible. We have used a coupled process of imagery and terrain analysis, yield maps, and electrical conductivity measurements to guide the locations of ERT surveys in this project (Tucker-Kulesza et al. 2017).more » « less
-
Phase estimation plays a central role in communications, sensing, and information processing. Quantum correlated states, such as squeezed states, enable phase estimation beyond the shot-noise limit, and in principle approach the ultimate quantum limit in precision, when paired with optimal quantum measurements. However, physical realizations of optimal quantum measurements for optical phase estimation with quantum-correlated states are still unknown. Here we address this problem by introducing an adaptive Gaussian measurement strategy for optical phase estimation with squeezed vacuum states that, by construction, approaches the quantum limit in precision. This strategy builds from a comprehensive set of locally optimal POVMs through rotations and homodyne measurements and uses the Adaptive Quantum State Estimation framework for optimizing the adaptive measurement process, which, under certain regularity conditions, guarantees asymptotic optimality for this quantum parameter estimation problem. As a result, the adaptive phase estimation strategy based on locally-optimal homodyne measurements achieves the quantum limit within the phase interval of . Furthermore, we generalize this strategy by including heterodyne measurements, enabling phase estimation across the full range of phases from , where squeezed vacuum allows for unambiguous phase encoding. Remarkably, for this phase interval, which is the maximum range of phases that can be encoded in squeezed vacuum, this estimation strategy maintains an asymptotic quantum-optimal performance, representing a significant advancement in quantum metrology.more » « less
