Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging
| Main Author: | |
|---|---|
| Language: | English |
| Published: |
University of Dayton / OhioLINK
2020
|
| Subjects: | |
| Online Access: | http://rave.ohiolink.edu/etdc/view?acc_num=dayton1607694391536891 |
| id |
ndltd-OhioLink-oai-etd.ohiolink.edu-dayton1607694391536891 |
|---|---|
| record_format |
oai_dc |
| spelling |
ndltd-OhioLink-oai-etd.ohiolink.edu-dayton16076943915368912021-08-03T07:16:40Z Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging Hoffmire, Matthew A. Electrical Engineering anisoplanatic optical turbulence mitigation atmospheric turbulence turbulence simulator deep learning convolutional neural network We present a novel deep learning approach for restoring images degraded by atmospheric optical turbulence. We consider the case of terrestrial imaging over long ranges with a wide field of view. This produces an anisoplanatic imaging scenario where the turbulence warping and blurring varies spatially across the image. The proposed turbulence mitigation (TM) method assumes that a sequence of short exposure images are acquired. A block matching algorithm for dewarping observed frames is applied and the resulting images are averaged. A convolutional neural network (CNN) is then employed to perform spatially adaptive restoration. Training the CNN is accomplished using simulated data from a fast simulation tool that mimics turbulence effects and is capable of producing a large amount of degraded imagery from ground truth imagery rapidly. Testing is done using independent data simulated with a different well-validated and pioneering anisoplanatic turbulence simulator. The anisoplanatic simulator is known to be very accurate in modeling turbulence. However, because it is much more computationally demanding, it is used here only for producing the limited amount of testing data needed. Our proposed TM method is evaluated in a number of experiments using quantitative metrics. The quantitative analysis is made possible by virtue of having ground truth imagery that is available with simulated data. A number of restored images are also provided for subjective evaluation. We demonstrate that the new TM method outperforms all of the benchmark methods in the scenarios tested in this thesis. 2020 English text University of Dayton / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=dayton1607694391536891 http://rave.ohiolink.edu/etdc/view?acc_num=dayton1607694391536891 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws. |
| collection |
NDLTD |
| language |
English |
| sources |
NDLTD |
| topic |
Electrical Engineering anisoplanatic optical turbulence mitigation atmospheric turbulence turbulence simulator deep learning convolutional neural network |
| spellingShingle |
Electrical Engineering anisoplanatic optical turbulence mitigation atmospheric turbulence turbulence simulator deep learning convolutional neural network Hoffmire, Matthew A. Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging |
| author |
Hoffmire, Matthew A. |
| author_facet |
Hoffmire, Matthew A. |
| author_sort |
Hoffmire, Matthew A. |
| title |
Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging |
| title_short |
Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging |
| title_full |
Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging |
| title_fullStr |
Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging |
| title_full_unstemmed |
Deep Learning for Anisoplanatic Optical Turbulence Mitigation in Long Range Imaging |
| title_sort |
deep learning for anisoplanatic optical turbulence mitigation in long range imaging |
| publisher |
University of Dayton / OhioLINK |
| publishDate |
2020 |
| url |
http://rave.ohiolink.edu/etdc/view?acc_num=dayton1607694391536891 |
| work_keys_str_mv |
AT hoffmirematthewa deeplearningforanisoplanaticopticalturbulencemitigationinlongrangeimaging |
| _version_ |
1719457960765161472 |