基于人類視覺機制和粒子濾波的紅外目標(biāo)跟蹤方法
發(fā)布時間：2018-08-26 15:34
【摘要】：隨著計算機、紅外技術(shù)的發(fā)展,對紅外目標(biāo)進行精確識別與跟蹤的需求不斷增長。然而,紅外目標(biāo)檢測與跟蹤過程中經(jīng)常遇到目標(biāo)弱小、背景復(fù)雜、信噪比過低等情況。同時目標(biāo)在運動過程中,可能發(fā)生的目標(biāo)灰度變化、背景灰度變化、目標(biāo)被遮擋或暫時丟失等情況也增加了檢測與跟蹤的難度。因此,針對復(fù)雜背景下紅外弱小目標(biāo)檢測與跟蹤方法的研究具有十分重要的意義。本文基于此,從人類視覺系統(tǒng)優(yōu)勢應(yīng)用的角度對紅外目標(biāo)檢測與跟蹤的方法進行多方面的思考,以期對該理論的發(fā)展以及人類視覺系統(tǒng)在紅外目標(biāo)檢測與跟蹤中的實際應(yīng)用提供有益借鑒。本文以復(fù)雜背景下紅外弱小目標(biāo)的檢測與跟蹤為研究對象,指出了現(xiàn)階段結(jié)合人類視覺系統(tǒng)進行紅外弱小目標(biāo)檢測與跟蹤的研究背景,以及針對該課題研究的理論意義和實際意義;通過分析國內(nèi)外相關(guān)問題的算法研究,指出現(xiàn)階段復(fù)雜背景下紅外弱小目標(biāo)的檢測與跟蹤方法存在的局限性,通過計算圖像局部視覺對比度和自適應(yīng)閾值判定改進了基于人類視覺對比機制的紅外弱小目標(biāo)檢測方法,同時通過對比實驗,證實了該目標(biāo)檢測方法具有兼顧檢測準(zhǔn)確率和實時性的良好性能;通過分析以灰度特征為目標(biāo)單一特征的粒子濾波紅外目標(biāo)跟蹤方法的不足,提出了基于人類視覺對比機制和粒子濾波的紅外目標(biāo)跟蹤方法,充分模擬人類視覺對比機制,提取目標(biāo)區(qū)域局部視覺對比度顯著圖為跟蹤的目標(biāo)特征,建立了“九宮格”式目標(biāo)模板,并通過對比實驗結(jié)果分析,驗證了該算法的魯棒性;通過分析固定模板、自定義周期更新模板以及即時更新模板這幾種傳統(tǒng)粒子濾波紅外目標(biāo)跟蹤模板更新方法的局限性,提出了基于人類視覺對比機制和粒子濾波的紅外目標(biāo)跟蹤方法,模擬人類視覺系統(tǒng)的學(xué)習(xí)和記憶機制,對候選模板進行學(xué)習(xí)、匹配和記憶并建立三維模板庫,通過對比實驗結(jié)果分析,驗證了該算法在目標(biāo)背景變化復(fù)雜、自身尺寸極小情況下的適用性;最后,對本文做出總結(jié)與展望。
[Abstract]:With the development of computer and infrared technology, the demand for accurate recognition and tracking of infrared targets is increasing. However, infrared target detection and tracking often encounter weak targets, complex background and low signal-to-noise ratio (SNR). At the same time, in the process of moving the target, the possible changes of the gray level of the target, the change of the background gray, the occlusion or the temporary loss of the target also increase the difficulty of detection and tracking. Therefore, it is of great significance to study the detection and tracking methods of infrared dim targets in complex background. Based on this, this paper discusses the methods of infrared target detection and tracking from the point of view of the superiority of human visual system. In order to provide useful reference for the development of this theory and the practical application of human vision system in infrared target detection and tracking. In this paper, the detection and tracking of infrared dim and weak targets under complex background is studied, and the research background of infrared dim and weak target detection and tracking based on human vision system is pointed out. And the theoretical and practical significance of the research, through the analysis of the relevant problems at home and abroad algorithm research, pointed out the current complex background of infrared small and weak target detection and tracking methods have limitations. The infrared dim target detection method based on human visual contrast mechanism is improved by calculating the local visual contrast and adaptive threshold decision of image. At the same time, the contrast experiment is carried out. It is proved that the target detection method has good performance of both accurate and real-time detection, and the shortcomings of particle filter infrared target tracking method with grayscale feature as single feature are analyzed. An infrared target tracking method based on human visual contrast mechanism and particle filter is proposed, which fully simulates the human visual contrast mechanism, and extracts the salient visual contrast map of the target region as the target feature. The "nine-cell" target template is established, and the robustness of the algorithm is verified by comparing the experimental results, and by analyzing the fixed template, the robustness of the algorithm is verified. The limitation of several traditional particle filter infrared target tracking template updating methods such as custom periodic update template and instant update template is introduced. A new infrared target tracking method based on human visual contrast mechanism and particle filter is proposed. The learning and memory mechanism of human visual system is simulated, the candidate template is studied, matched and memorized, and the 3D template library is built. By comparing the experimental results, it is proved that the algorithm is complex in the change of target background. The applicability of this paper in the case of minimal size; finally, this paper is summarized and prospected.
【學(xué)位授予單位】：江蘇科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP391.41;TN219
本文編號：2205339
[Abstract]:With the development of computer and infrared technology, the demand for accurate recognition and tracking of infrared targets is increasing. However, infrared target detection and tracking often encounter weak targets, complex background and low signal-to-noise ratio (SNR). At the same time, in the process of moving the target, the possible changes of the gray level of the target, the change of the background gray, the occlusion or the temporary loss of the target also increase the difficulty of detection and tracking. Therefore, it is of great significance to study the detection and tracking methods of infrared dim targets in complex background. Based on this, this paper discusses the methods of infrared target detection and tracking from the point of view of the superiority of human visual system. In order to provide useful reference for the development of this theory and the practical application of human vision system in infrared target detection and tracking. In this paper, the detection and tracking of infrared dim and weak targets under complex background is studied, and the research background of infrared dim and weak target detection and tracking based on human vision system is pointed out. And the theoretical and practical significance of the research, through the analysis of the relevant problems at home and abroad algorithm research, pointed out the current complex background of infrared small and weak target detection and tracking methods have limitations. The infrared dim target detection method based on human visual contrast mechanism is improved by calculating the local visual contrast and adaptive threshold decision of image. At the same time, the contrast experiment is carried out. It is proved that the target detection method has good performance of both accurate and real-time detection, and the shortcomings of particle filter infrared target tracking method with grayscale feature as single feature are analyzed. An infrared target tracking method based on human visual contrast mechanism and particle filter is proposed, which fully simulates the human visual contrast mechanism, and extracts the salient visual contrast map of the target region as the target feature. The "nine-cell" target template is established, and the robustness of the algorithm is verified by comparing the experimental results, and by analyzing the fixed template, the robustness of the algorithm is verified. The limitation of several traditional particle filter infrared target tracking template updating methods such as custom periodic update template and instant update template is introduced. A new infrared target tracking method based on human visual contrast mechanism and particle filter is proposed. The learning and memory mechanism of human visual system is simulated, the candidate template is studied, matched and memorized, and the 3D template library is built. By comparing the experimental results, it is proved that the algorithm is complex in the change of target background. The applicability of this paper in the case of minimal size; finally, this paper is summarized and prospected.
【學(xué)位授予單位】：江蘇科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP391.41;TN219
【參考文獻】
相關(guān)期刊論文 前5條
1 侯旺;孫曉亮;尚洋;于起峰;;紅外弱小目標(biāo)檢測技術(shù)研究現(xiàn)狀與發(fā)展趨勢[J];紅外技術(shù);2015年01期
2 張寅生;龍云利;劉琨;;Track-before-detect algorithm for maneuvering infrared weak multiple targets via particle filter[J];Chinese Optics Letters;2014年10期
3 朱明清;王智靈;陳宗海;;基于人類視覺智能和粒子濾波的魯棒目標(biāo)跟蹤算法[J];控制與決策;2012年11期
4 萬九卿;梁旭;馬志峰;;基于自適應(yīng)觀測模型交互多模型粒子濾波的紅外機動目標(biāo)跟蹤[J];電子學(xué)報;2011年03期
5 李金;胡文廣;;基于顏色的快速人體跟蹤及遮擋處理[J];智能系統(tǒng)學(xué)報;2010年04期
，本文編號：2205339
本文鏈接：http://www.sikaile.net/kejilunwen/dianzigongchenglunwen/2205339.html
最近更新
教材專著
