基于視覺注意機制的UWB SAR葉簇隱蔽目標變化檢測技術(shù)研究
發(fā)布時間：2019-05-27 16:48
【摘要】：低頻超寬帶合成孔徑雷達(Uitra Wide Band Synthetic Aperture Radar,UWB SAR)因其強葉簇穿透能力,逐步在民事、軍事領(lǐng)域受到廣泛應(yīng)用。然而,隨著UWB SAR系統(tǒng)分辨率提高,應(yīng)用范圍的不斷擴大,對于UWB SAR葉簇隱蔽目標快速、精準檢測的需求日趨強烈。因此本文從人類視覺系統(tǒng)的視覺注意機制入手,分析并建立適用于低頻UWB SAR圖像的視覺注意模型,提出了一種無監(jiān)督、全自動、高效率的低頻UWB SAR葉簇隱蔽目標變化檢測方法,取得了良好的檢測效果。論文工作及創(chuàng)新點主要包括:1、視覺注意機制與圖像處理相結(jié)合的可行性分析。本文依據(jù)人類視覺系統(tǒng)的現(xiàn)有資料,從生物學(xué)角度入手,研究人類視覺系統(tǒng)工作機理,總結(jié)其信息處理的主要特點;探討將視覺注意機制應(yīng)用于低頻UWB SAR圖像處理過程的可行性,并構(gòu)建適用于低頻UWB SAR圖像處理的視覺注意模型,為后期結(jié)合視覺注意機制展開低頻UWB SAR變化檢測技術(shù)的研究提供理論基礎(chǔ)。2、低頻UWB SAR圖像數(shù)據(jù)預(yù)處理,即圖像可視化技術(shù)研究。首先結(jié)合人眼視覺特性,確定低頻UWB SAR圖像可視化的視覺依據(jù),分析低頻UWB SAR圖像內(nèi)在的數(shù)據(jù)統(tǒng)計特性和灰度分布模型,并針對低頻UWB SAR圖像的特點,提出符合視覺特性的低頻UWB SAR快速可視化算法。在同等條件下該方法與傳統(tǒng)的對數(shù)映射法、線性映射法相比,所得結(jié)果在等效視數(shù)、處理耗時等指標的考核中表現(xiàn)優(yōu)異,更適于人眼判讀,且能夠滿足低頻UWB SAR系統(tǒng)實時處理等方面的應(yīng)用需求。3、基于視覺注意機制展開多時相UWB SAR葉簇隱蔽目標變化檢測技術(shù)研究。首先從視覺分析角度出發(fā),模擬人類視覺系統(tǒng),制構(gòu)建適用于圖像分析的高斯金字塔模型,根據(jù)人類視覺系統(tǒng)的生理結(jié)構(gòu)與認知特點,提出基于視覺注意機制的UWB SAR葉簇隱蔽目標變化檢測算法,該算法將復(fù)雜圖像簡化為視覺焦點集合,并利用圖像局部鄰域信息和目標的空間相關(guān)特性對視覺注意焦點進行分層篩選,以提高變化檢測精度。與廣泛應(yīng)用的基于CFAR的變化檢測技術(shù)相比,同等情況下,基于視覺注意機制的UWB SAR葉簇隱蔽目標檢測算法不但能有效彌補CFAR檢測方法存在的漏檢問題,而且處理耗時量僅為CFAR檢測的12%,因而在實際應(yīng)用中具有重大的應(yīng)用前景。
[Abstract]:Low frequency ultra-broadband synthetic aperture radar (Uitra Wide Band Synthetic Aperture Radar,UWB SAR) has been widely used in civil and military fields because of its strong blade cluster penetration ability. However, with the improvement of the resolution of UWB SAR system and the continuous expansion of its application range, the demand for fast hidden targets and accurate detection of UWB SAR leaf clusters is becoming more and more strong. Therefore, starting with the visual attention mechanism of human visual system, this paper analyzes and establishes a visual attention model suitable for low frequency UWB SAR images, and proposes an unsupervised, fully automatic and efficient method for detecting the change of hidden targets in low frequency UWB SAR leaf clusters. Good detection results have been obtained. The main work and innovations of this paper are as follows: 1. The feasibility analysis of the combination of visual attention mechanism and image processing. Based on the existing data of human visual system, this paper studies the working mechanism of human visual system from the biological point of view, and summarizes the main characteristics of its information processing. This paper discusses the feasibility of applying visual attention mechanism to low frequency UWB SAR image processing, and constructs a visual attention model suitable for low frequency UWB SAR image processing. It provides a theoretical basis for the later research of low frequency UWB SAR change detection technology combined with visual attention mechanism. 2, low frequency UWB SAR image data preprocessing, that is, image visualization technology research. Firstly, combined with the human visual characteristics, the visual basis of low frequency UWB SAR image visualization is determined, and the inherent data statistical characteristics and gray distribution model of low frequency UWB SAR image are analyzed, and the characteristics of low frequency UWB SAR image are analyzed. A fast visualization algorithm for low frequency UWB SAR is proposed, which accords with visual characteristics. Under the same conditions, compared with the traditional logarithmic mapping method and linear mapping method, the results obtained by this method have excellent performance in the assessment of equivalent visual number and processing time, and are more suitable for human eye interpretation. And can meet the application requirements of low frequency UWB SAR system real-time processing. 3. Based on the visual attention mechanism, the multi-temporal UWB SAR leaf cluster hidden target change detection technology is studied. First of all, from the point of view of visual analysis, the human visual system is simulated, and the Gao Si pyramid model suitable for image analysis is constructed. According to the physiological structure and cognitive characteristics of human visual system, A change detection algorithm for hidden targets in UWB SAR leaf clusters based on visual attention mechanism is proposed, which simplifies complex images to visual focus sets. In order to improve the accuracy of change detection, the visual attention focus is screened layered by using the local neighborhood information of the image and the spatial correlation characteristics of the target. Compared with the widely used change detection technology based on CFAR, the hidden target detection algorithm of UWB SAR leaf cluster based on visual attention mechanism can not only effectively make up for the missed detection problem of CFAR detection method. Moreover, the processing time consumption is only 12% of that of CFAR detection, so it has a great application prospect in practical application.
【學(xué)位授予單位】：國防科學(xué)技術(shù)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TN958
本文編號：2486311
[Abstract]:Low frequency ultra-broadband synthetic aperture radar (Uitra Wide Band Synthetic Aperture Radar,UWB SAR) has been widely used in civil and military fields because of its strong blade cluster penetration ability. However, with the improvement of the resolution of UWB SAR system and the continuous expansion of its application range, the demand for fast hidden targets and accurate detection of UWB SAR leaf clusters is becoming more and more strong. Therefore, starting with the visual attention mechanism of human visual system, this paper analyzes and establishes a visual attention model suitable for low frequency UWB SAR images, and proposes an unsupervised, fully automatic and efficient method for detecting the change of hidden targets in low frequency UWB SAR leaf clusters. Good detection results have been obtained. The main work and innovations of this paper are as follows: 1. The feasibility analysis of the combination of visual attention mechanism and image processing. Based on the existing data of human visual system, this paper studies the working mechanism of human visual system from the biological point of view, and summarizes the main characteristics of its information processing. This paper discusses the feasibility of applying visual attention mechanism to low frequency UWB SAR image processing, and constructs a visual attention model suitable for low frequency UWB SAR image processing. It provides a theoretical basis for the later research of low frequency UWB SAR change detection technology combined with visual attention mechanism. 2, low frequency UWB SAR image data preprocessing, that is, image visualization technology research. Firstly, combined with the human visual characteristics, the visual basis of low frequency UWB SAR image visualization is determined, and the inherent data statistical characteristics and gray distribution model of low frequency UWB SAR image are analyzed, and the characteristics of low frequency UWB SAR image are analyzed. A fast visualization algorithm for low frequency UWB SAR is proposed, which accords with visual characteristics. Under the same conditions, compared with the traditional logarithmic mapping method and linear mapping method, the results obtained by this method have excellent performance in the assessment of equivalent visual number and processing time, and are more suitable for human eye interpretation. And can meet the application requirements of low frequency UWB SAR system real-time processing. 3. Based on the visual attention mechanism, the multi-temporal UWB SAR leaf cluster hidden target change detection technology is studied. First of all, from the point of view of visual analysis, the human visual system is simulated, and the Gao Si pyramid model suitable for image analysis is constructed. According to the physiological structure and cognitive characteristics of human visual system, A change detection algorithm for hidden targets in UWB SAR leaf clusters based on visual attention mechanism is proposed, which simplifies complex images to visual focus sets. In order to improve the accuracy of change detection, the visual attention focus is screened layered by using the local neighborhood information of the image and the spatial correlation characteristics of the target. Compared with the widely used change detection technology based on CFAR, the hidden target detection algorithm of UWB SAR leaf cluster based on visual attention mechanism can not only effectively make up for the missed detection problem of CFAR detection method. Moreover, the processing time consumption is only 12% of that of CFAR detection, so it has a great application prospect in practical application.
【學(xué)位授予單位】：國防科學(xué)技術(shù)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TN958
【參考文獻】
相關(guān)期刊論文 前5條
1 劉娟妮;彭進業(yè);李大湘;王平;;基于譜殘差和多分辨率分析的顯著目標檢測[J];中國圖象圖形學(xué)報;2011年02期
2 王廣學(xué);黃曉濤;周智敏;;基于圖像分割的VHF SAR葉簇隱蔽目標差值變化檢測[J];電子學(xué)報;2010年09期
3 周旭;保錚;;SAR目標特性分析技術(shù)[J];計算機工程與科學(xué);2008年07期
4 張鵬,王潤生;基于視點轉(zhuǎn)移和視區(qū)追蹤的圖像顯著區(qū)域檢測[J];軟件學(xué)報;2004年06期
5 徐孟俠;圖像編碼的進展[J];通信學(xué)報;1993年02期
相關(guān)博士學(xué)位論文 前1條
1 鐘家強;基于多時相遙感圖像的變化檢測[D];國防科學(xué)技術(shù)大學(xué);2005年
，本文編號：2486311
本文鏈接：http://sikaile.net/kejilunwen/wltx/2486311.html
最近更新
教材專著
