【摘要】：基于運(yùn)動(dòng)平臺(tái)的紅外目標(biāo)檢測(cè)是紅外預(yù)警系統(tǒng)信息處理的核心技術(shù)。單幀圖像的背景抑制是前端信息處理的關(guān)鍵環(huán)節(jié),序列圖像的幀間圖像配準(zhǔn)是確定目標(biāo)軌跡不容忽視的環(huán)節(jié),對(duì)目標(biāo)檢測(cè)性能都有重要的影響。本文針對(duì)上述兩個(gè)預(yù)處理技術(shù)進(jìn)行深入研究,并取得了一定的成果。對(duì)單幀圖像的背景抑制方法的研究方面,主要闡述了全變分理論的概念及其發(fā)展;推導(dǎo)全變分背景抑制算法的原理,在A.Chambolle投影算法的基礎(chǔ)上改進(jìn)后應(yīng)用于單幀紅外圖像的背景抑制;實(shí)驗(yàn)仿真不僅驗(yàn)證基于全變分理論的改進(jìn)算法對(duì)紅外圖像背景抑制的性能,而且與三種經(jīng)典背景抑制算法進(jìn)行比較,彰顯改進(jìn)算法的特點(diǎn)和優(yōu)勢(shì)。對(duì)序列圖像的幀間配準(zhǔn)技術(shù)研究方面,主要論述了SIFT算法的基本原理和主要步驟,并針對(duì)特征點(diǎn)提取和特征點(diǎn)匹配兩個(gè)重點(diǎn)環(huán)節(jié)進(jìn)行改進(jìn)。多尺度特征提取算法存在邊緣響應(yīng)影響,采用Canny算子進(jìn)行改進(jìn);在原匹配算法的基礎(chǔ)上,改進(jìn)初匹配效率,結(jié)合馬氏距離進(jìn)行精匹配,并采用RANSAC算法進(jìn)行變換參數(shù)估計(jì)。最后分別針對(duì)改進(jìn)算法進(jìn)行仿真實(shí)驗(yàn),驗(yàn)證其性能。論文研究的基于全變分理論的單幀圖像背景抑制算法和基于SIFT算法的序列圖像幀間配準(zhǔn)算法都能夠應(yīng)用于紅外目標(biāo)檢測(cè),不但能夠解決復(fù)雜背景的問(wèn)題,還能夠充分考慮和解決運(yùn)動(dòng)平臺(tái)的影響,對(duì)紅外空間預(yù)警技術(shù)的發(fā)展提供一定的支持。
[Abstract]:Infrared target detection based on moving platform is the core technology of infrared early warning system information processing. Background suppression of single frame image is a key step in front-end information processing. Image registration between frames of sequence images can not be ignored in determining the track of target, which has an important impact on the performance of target detection. In this paper, the above two pretreatment techniques are studied, and some achievements have been made. In the research of background suppression method of single frame image, the concept and development of total variational theory are expounded, and the principle of total variation background suppression algorithm is deduced. Based on the A. Chambolle projection algorithm, it is applied to the background suppression of a single infrared image, and the performance of the improved algorithm based on the total variational theory is verified by the experimental simulation. And compared with the three classical background suppression algorithms, the improved algorithm shows the characteristics and advantages. In this paper, the basic principle and main steps of sift algorithm are discussed, and the two key links of feature point extraction and feature point matching are improved. The multi-scale feature extraction algorithm has the influence of edge response, which is improved by Canny operator. Based on the original matching algorithm, the initial matching efficiency is improved, and the Mahalanobis distance is combined to perform the precise matching. The RANSAC algorithm is used to estimate the transformation parameters. Finally, the improved algorithm is simulated to verify its performance. The single frame background suppression algorithm based on total variational theory and the sequential image frame registration algorithm based on sift algorithm can be applied to infrared target detection, which can not only solve the problem of complex background. It can also fully consider and solve the influence of motion platform, and provide some support for the development of infrared space early warning technology.
【學(xué)位授予單位】：國(guó)防科學(xué)技術(shù)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2015
【分類號(hào)】：TP391.41;TN219

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 李小萍;盛景泰;;基于空中運(yùn)動(dòng)平臺(tái)的雷達(dá)任務(wù)系統(tǒng)可靠性研究[J];電子產(chǎn)品可靠性與環(huán)境試驗(yàn);2011年03期

2 董永謙;狄希遠(yuǎn);張志耀;田芳;;高速精密運(yùn)動(dòng)平臺(tái)研制[J];電子工藝技術(shù);2011年03期

3 楊國(guó)哲,王立平,郁鼎文,黃歡;三坐標(biāo)精密運(yùn)動(dòng)平臺(tái)底座的設(shè)計(jì)[J];機(jī)械工程師;2005年07期

4 王貴平;董永謙;田芳;狄希遠(yuǎn);;生瓷帶打孔機(jī)運(yùn)動(dòng)平臺(tái)的設(shè)計(jì)[J];電子工藝技術(shù);2009年02期

5 紀(jì)大壯;楊小雷;王國(guó)宏;李勝軍;詹毅;;基于改進(jìn)粒子濾波的運(yùn)動(dòng)平臺(tái)定位算法[J];系統(tǒng)仿真學(xué)報(bào);2010年09期

6 楊國(guó)哲;王立平;郁鼎文;崔巍;;三維微細(xì)電火花加工運(yùn)動(dòng)平臺(tái)的誤差模型建立[J];機(jī)械工程師;2006年06期

7 孫浩;王程;王潤(rùn)生;;基于多傳感器融合的運(yùn)動(dòng)平臺(tái)運(yùn)動(dòng)目標(biāo)檢測(cè)[J];計(jì)算機(jī)應(yīng)用;2008年04期

8 孫偉;張科;李言俊;姜偉;;運(yùn)動(dòng)平臺(tái)上基于粒子濾波的目標(biāo)跟蹤技術(shù)[J];航空計(jì)算技術(shù);2010年05期

9 楊一博;尹文生;朱煜;王春洪;;粗精動(dòng)超精密運(yùn)動(dòng)平臺(tái)系統(tǒng)建模與分析研究[J];中國(guó)機(jī)械工程;2008年23期

10 趙岐誠(chéng);江晶;蓋旭剛;;空中多運(yùn)動(dòng)平臺(tái)多傳感器目標(biāo)跟蹤[J];現(xiàn)代雷達(dá);2009年10期

相關(guān)會(huì)議論文 前8條

1 代小林;叢大成;韓俊偉;李洪人;;基于凱恩方程的并聯(lián)運(yùn)動(dòng)平臺(tái)多剛體動(dòng)力學(xué)建模[A];第五屆全國(guó)流體傳動(dòng)與控制學(xué)術(shù)會(huì)議暨2008年中國(guó)航空學(xué)會(huì)液壓與氣動(dòng)學(xué)術(shù)會(huì)議論文集[C];2008年

2 由俊生;由勇;許敘遙;馬月欣;;持續(xù)載荷飛行仿真運(yùn)動(dòng)平臺(tái)的預(yù)測(cè)控制研究[A];中國(guó)航空學(xué)會(huì)控制與應(yīng)用第十二屆學(xué)術(shù)年會(huì)論文集[C];2006年

3 韓e，

本文編號(hào)：2122134