本文選題：危險(xiǎn)品檢測(cè)　切入點(diǎn)：FastICA算法　出處：《成都理工大學(xué)》2016年碩士論文

【摘要】：近年來,國(guó)際形勢(shì)動(dòng)蕩,恐怖襲擊事件時(shí)有發(fā)生。公共安全問題日趨嚴(yán)重,對(duì)于公交車上發(fā)生的危害公共安全的事件也越來越多,犯罪分子常常采用汽油、管制刀具等對(duì)普通民眾進(jìn)行傷害,造成的社會(huì)影響極其惡劣。因此為了在檢測(cè)到的復(fù)雜混合信號(hào)中,將危險(xiǎn)品信號(hào)分離出來,研究一種有效的盲源分離算法是非常重要的。盲源分離研究的是在不知道源信號(hào)和將其混合的方式情況下,僅通過源信號(hào)的一些統(tǒng)計(jì)特性,然后根據(jù)觀測(cè)信號(hào)(混合信號(hào))來恢復(fù)出源信號(hào)。獨(dú)立分量分析是盲源分離方法中的一種特殊算法,它是建立在源信號(hào)相互獨(dú)立情況下的一種算法。而快速獨(dú)立分量分析(FastICA)算法是利用了最大化信號(hào)的非高斯性,該算法采用了牛頓方法去處理計(jì)算值的多個(gè)采樣數(shù)據(jù),每步的迭代計(jì)算能從混合信號(hào)中間分離出一個(gè)獨(dú)立的成分,屬于獨(dú)立分量分析算法中的一種快速計(jì)算方法。在本論文中,首先分析了危險(xiǎn)品微波檢測(cè)的原理,根據(jù)檢測(cè)的原理對(duì)檢測(cè)系統(tǒng)的功能進(jìn)行了實(shí)現(xiàn)。其中以汽油、乙醇等常見危險(xiǎn)品的介電常數(shù)為基礎(chǔ),遠(yuǎn)距離非接觸式測(cè)量危險(xiǎn)品S參數(shù),并對(duì)S參數(shù)進(jìn)行分析,實(shí)現(xiàn)利用厘米波遠(yuǎn)距離非接觸式地排查危險(xiǎn)品。然后對(duì)算法進(jìn)行了研究,針對(duì)不同的情況使用了兩種算:累加算法和Fast ICA算法。在只檢測(cè)到一種危險(xiǎn)品的時(shí)候,我們只需要判斷出它是什么危險(xiǎn)品,論文中分別介紹了累積求和與累加算法來進(jìn)行判斷,最終根據(jù)得出的結(jié)果判斷出累加算法比累積求和的算法更加適用于危險(xiǎn)品檢測(cè)中;對(duì)于同時(shí)檢測(cè)到多種危險(xiǎn)品信號(hào)的時(shí)候,由于此時(shí)的信號(hào)是一個(gè)觀測(cè)信號(hào),因而我們需要利用Fast ICA算法對(duì)其進(jìn)行盲源分離,得到不同危險(xiǎn)品的獨(dú)立信號(hào)。在論文中分別以水和空氣作為參考,通過對(duì)比發(fā)現(xiàn)以水為參考得出的解混信號(hào)與源信號(hào)的相似程度要比空氣的要加接近。為了將危險(xiǎn)品檢測(cè)的算法進(jìn)行實(shí)現(xiàn),設(shè)計(jì)了一個(gè)危險(xiǎn)品檢測(cè)軟件系統(tǒng),系統(tǒng)包括上位機(jī)部分與下位機(jī)部分。開發(fā)危險(xiǎn)品檢測(cè)軟件系統(tǒng)時(shí),對(duì)現(xiàn)有的一些可以開發(fā)界面的工具進(jìn)行了一個(gè)簡(jiǎn)單的對(duì)比,最終選擇了Qt開發(fā)框架。在通信方面,采用高可靠性TCP協(xié)議的客戶端/服務(wù)器模式來進(jìn)行通信,在下位機(jī)請(qǐng)求連接上位機(jī)并連接成功時(shí),就可以開始下位機(jī)信息的發(fā)送和上位機(jī)數(shù)據(jù)的接收。當(dāng)上位機(jī)接收檢測(cè)到危險(xiǎn)品的信息后,就會(huì)在界面上顯示危險(xiǎn)品圖片和對(duì)圖片的說明,發(fā)出報(bào)警聲,并且還能將接收到的信息進(jìn)行截圖保存。在危險(xiǎn)品檢測(cè)算法的研究與實(shí)現(xiàn)的過程中,對(duì)算法和程序進(jìn)行了不斷的改進(jìn)和調(diào)試,改正了各種細(xì)節(jié)錯(cuò)誤。初步實(shí)現(xiàn)了能夠?qū)z測(cè)到的危險(xiǎn)品進(jìn)行判別,并通過危險(xiǎn)品檢測(cè)軟件系統(tǒng)來實(shí)現(xiàn)報(bào)警,在通信的穩(wěn)定性以及上位機(jī)和下位機(jī)各自的功能等方面都基本達(dá)到了要求。
[Abstract]:In recent years, the international situation has been turbulent, terrorist attacks have occurred from time to time. Public safety problems are becoming more and more serious, and there are more and more incidents that endanger public safety on buses. Criminals often use gasoline. In order to separate dangerous signals from the complex mixed signals detected, the social impact caused by the harm done to ordinary people by controlling knives and others is extremely bad. It is very important to study an effective blind source separation algorithm. Blind source separation studies only some statistical characteristics of the source signal without knowing the source signal and mixing it. Independent component analysis (ICA) is a special algorithm in blind source separation. The fast independent component analysis (FastICA) algorithm takes advantage of the non-#china_person0# property of the maximized signal, and the Newton method is used to deal with multiple sampled data of the calculated value, the algorithm is based on the source signal being independent of each other, and the fast independent component analysis (FastICA) algorithm takes advantage of the non-#china_person0# property of the maximized signal. The iterative calculation of each step can separate an independent component from the mixed signal, which is a fast calculation method in the independent component analysis algorithm. In this paper, the principle of microwave detection of dangerous goods is first analyzed. According to the principle of detection, the function of the detection system is realized. Based on the dielectric constant of gasoline, ethanol and other common dangerous goods, the S parameters of dangerous goods are measured and analyzed by remote non-contact method. In this paper, centimeter wave is used to detect dangerous goods in a long distance and non-contact way. Then, the algorithm is studied, and two kinds of algorithms are used for different cases: accumulative algorithm and Fast ICA algorithm. When only one kind of dangerous goods is detected, when only one kind of dangerous goods is detected, We only need to judge what dangerous goods it is. In this paper, the cumulative summation and accumulation algorithms are introduced to judge. Finally, it is concluded that the accumulative algorithm is more suitable for dangerous goods detection than the cumulative summation algorithm. When a variety of dangerous goods signals are detected at the same time, because the signal is an observational signal, we need to use the Fast ICA algorithm to separate its blind sources. Get the independent signal of different dangerous goods. In the paper, water and air are respectively used as reference, It is found by comparison that the similarity between the decontamination signal and the source signal obtained by reference to water is closer than that of air. In order to realize the algorithm of dangerous goods detection, a dangerous goods detection software system is designed. The system includes the upper computer part and the lower computer part. When developing the dangerous goods detection software system, the paper makes a simple comparison to some existing tools that can develop interface, and finally chooses the QT development framework. The client / server mode of high reliability TCP protocol is used to communicate. When the lower computer requests to connect with the host computer and the connection is successful, When the host computer receives the information of the dangerous goods, it will display the picture of dangerous goods and explain the picture on the interface, and issue the alarm sound. In the process of research and implementation of dangerous goods detection algorithm, the algorithm and program have been continuously improved and debugged. Has corrected all kinds of details error. Has realized preliminarily can detect the dangerous goods to carry on the discrimination, and through the dangerous goods detection software system realizes the alarm, The stability of communication and the functions of upper computer and lower computer have basically met the requirements.
【學(xué)位授予單位】：成都理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：D63;TN911.7

【相似文獻(xiàn)】

相關(guān)會(huì)議論文 前2條

1 王鳳琴;樊養(yǎng)余;婁慧敏;;一種實(shí)時(shí)檢測(cè)QRS波的算法及程序設(shè)計(jì)[A];信號(hào)與信息處理技術(shù)第三屆信號(hào)與信息處理全國(guó)聯(lián)合學(xué)術(shù)會(huì)議論文集[C];2004年

2 譚鵬程;洪偉榮;;過程動(dòng)態(tài)優(yōu)化的內(nèi)點(diǎn)擬序貫算法[A];PCC2009—第20屆中國(guó)過程控制會(huì)議論文集[C];2009年

相關(guān)博士學(xué)位論文 前3條

1 趙進(jìn)慧;膜計(jì)算仿生優(yōu)化算法及應(yīng)用研究[D];浙江大學(xué);2010年

2 劉小龍;細(xì)菌覓食優(yōu)化算法的改進(jìn)及應(yīng)用[D];華南理工大學(xué);2011年

3 高尚;蟻群算法理論、應(yīng)用及其與其它算法的混合[D];南京理工大學(xué);2005年

相關(guān)碩士學(xué)位論文 前10條

1 向宏衛(wèi);面向多核向量處理器的FFT算法設(shè)計(jì)與實(shí)現(xiàn)[D];國(guó)防科學(xué)技術(shù)大學(xué);2014年

2 吳清泉;基于DM6467T視頻圖像去霧算法研究與實(shí)現(xiàn)[D];中國(guó)科學(xué)院研究生院(西安光學(xué)精密機(jī)械研究所);2015年

3 王文華;交通分配PAS算法的研究與應(yīng)用[D];東南大學(xué);2015年

4 丁召杰;RSA算法之64位快速實(shí)現(xiàn)[D];山東大學(xué);2016年

5 任作琳;風(fēng)驅(qū)動(dòng)優(yōu)化算法及其在電磁綜合問題中的應(yīng)用研究[D];江蘇科技大學(xué);2016年

6 陳春梅;危險(xiǎn)品檢測(cè)算法的研究與實(shí)現(xiàn)[D];成都理工大學(xué);2016年

7 趙賓賓;進(jìn)化算法研究平臺(tái)的設(shè)計(jì)與開發(fā)[D];太原科技大學(xué);2012年

8 王崇寶;蟻群算法熵收斂性分析與應(yīng)用[D];四川師范大學(xué);2009年

9 李春輝;在CAI中算法的逆向演示[D];西安科技大學(xué);2009年

10 李強(qiáng);基于控件技術(shù)的算法可視化的研究[D];太原理工大學(xué);2007年

，

本文編號(hào)：1696997