GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images
碩士 === 國立臺北科技大學 === 電機工程系所 === 105 === Color encoding or assignment of multi-polarization or fully polarimetric synthetic aperture radar (PolSAR) image is vital for visual display and interpretation of the polarimetric information. It is a common practice to use RGB or HIS color space to display t...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2017
|
Online Access: | http://ndltd.ncl.edu.tw/handle/4b52y9 |
id |
ndltd-TW-105TIT05442123 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-105TIT054421232019-05-15T23:53:43Z http://ndltd.ncl.edu.tw/handle/4b52y9 GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images 以 GPU實現高效能CIE-Lab 色彩空間轉換應用於全極化合成孔徑雷達影像 Zheng-Hui Lin 林政輝 碩士 國立臺北科技大學 電機工程系所 105 Color encoding or assignment of multi-polarization or fully polarimetric synthetic aperture radar (PolSAR) image is vital for visual display and interpretation of the polarimetric information. It is a common practice to use RGB or HIS color space to display the chromatic information for polarization-encoded model-based target decomposition of PolSAR images. However, to express the multi partial Pol SAR’s chroma, the basic RGB area is not cover all the perception system of human. Our research used the frame of color distribution which based on the evenly CIE-Lab color space, for better visual aware and information explanation. Let color have strongly visual explanation which through every kinds of parameter. The new color distribution not only keep the pole feature’s color but also enhance the total power of target’s information. But it spends much time to converse the color space from PolSAR image. Thus, we adopt the Compute Unified Device Architecture (CUDA) which is parallel structure of Graphics Processing Units (GPU) to speed up the transferring-time. The experimental result shows that we do the data’s conversion with parallel structure not only can be more efficient on the running time than using CPU but get the correct color space. 于治平 張陽郎 2017 學位論文 ; thesis 33 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立臺北科技大學 === 電機工程系所 === 105 === Color encoding or assignment of multi-polarization or fully polarimetric synthetic aperture radar (PolSAR) image is vital for visual display and interpretation of the polarimetric information. It is a common practice to use RGB or HIS color space to display the chromatic information for polarization-encoded model-based target decomposition of PolSAR images. However, to express the multi partial Pol SAR’s chroma, the basic RGB area is not cover all the perception system of human.
Our research used the frame of color distribution which based on the evenly CIE-Lab color space, for better visual aware and information explanation. Let color have strongly visual explanation which through every kinds of parameter. The new color distribution not only keep the pole feature’s color but also enhance the total power of target’s information. But it spends much time to converse the color space from PolSAR image. Thus, we adopt the Compute Unified Device Architecture (CUDA) which is parallel structure of Graphics Processing Units (GPU) to speed up the transferring-time.
The experimental result shows that we do the data’s conversion with parallel structure not only can be more efficient on the running time than using CPU but get the correct color space.
|
author2 |
于治平 |
author_facet |
于治平 Zheng-Hui Lin 林政輝 |
author |
Zheng-Hui Lin 林政輝 |
spellingShingle |
Zheng-Hui Lin 林政輝 GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images |
author_sort |
Zheng-Hui Lin |
title |
GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images |
title_short |
GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images |
title_full |
GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images |
title_fullStr |
GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images |
title_full_unstemmed |
GPU-Acceleration of High Performance CIE-Lab Color Space Conversion for Fully Polarimetric Synthetic Aperture Radar Images |
title_sort |
gpu-acceleration of high performance cie-lab color space conversion for fully polarimetric synthetic aperture radar images |
publishDate |
2017 |
url |
http://ndltd.ncl.edu.tw/handle/4b52y9 |
work_keys_str_mv |
AT zhenghuilin gpuaccelerationofhighperformancecielabcolorspaceconversionforfullypolarimetricsyntheticapertureradarimages AT línzhènghuī gpuaccelerationofhighperformancecielabcolorspaceconversionforfullypolarimetricsyntheticapertureradarimages AT zhenghuilin yǐgpushíxiàngāoxiàonéngcielabsècǎikōngjiānzhuǎnhuànyīngyòngyúquánjíhuàhéchéngkǒngjìngléidáyǐngxiàng AT línzhènghuī yǐgpushíxiàngāoxiàonéngcielabsècǎikōngjiānzhuǎnhuànyīngyòngyúquánjíhuàhéchéngkǒngjìngléidáyǐngxiàng |
_version_ |
1719156347243593728 |