本文選題：交通標(biāo)志 + 霧天圖像��； 參考：《天津大學(xué)》2016年碩士論文

【摘要】：伴隨著智能交通系統(tǒng)的迅速發(fā)展,道路交通標(biāo)志識(shí)別作為智能交通系統(tǒng)的重要組成部分,逐漸成為當(dāng)前研究熱點(diǎn)。然而,在我國(guó)華北和東北地區(qū)經(jīng)常出現(xiàn)的霧霾等惡劣天氣,使現(xiàn)有的交通標(biāo)志檢測(cè)識(shí)別率嚴(yán)重降低,無(wú)法滿(mǎn)足現(xiàn)實(shí)需求。為了提高霧霾天氣下交通標(biāo)志的檢測(cè)識(shí)別率,本文設(shè)計(jì)了一套交通標(biāo)志檢測(cè)識(shí)別系統(tǒng)。首先,研究霧霾天氣下標(biāo)志識(shí)別系統(tǒng)的第一個(gè)環(huán)節(jié)——霧霾檢測(cè)。通過(guò)研究大量同一場(chǎng)景清晰圖像與霧霾圖像的區(qū)別,得出霧霾圖像的四個(gè)特征:圖像亮度、圖像對(duì)比度、圖像辨識(shí)度、霧霾區(qū)域單一性,設(shè)計(jì)了一個(gè)基于圖像亮度和圖像對(duì)比度兩個(gè)特征的Fisher分類(lèi)器,來(lái)判別當(dāng)前圖像是否為霧霾圖像。其次,圖像被認(rèn)證為霧霾圖像后,需進(jìn)行清晰化復(fù)原。霧天圖像清晰化技術(shù)研究主要包括兩個(gè)方向:一個(gè)是基于數(shù)學(xué)模型的圖像復(fù)原技術(shù),另一個(gè)是基于人體視覺(jué)的圖像增強(qiáng)技術(shù)。概述了兩個(gè)研究方向的經(jīng)典算法,并完成基于MATLAB平臺(tái)的試驗(yàn)仿真,總結(jié)仿真結(jié)果。文章采用圖像復(fù)原技術(shù),建立大氣光衰減模型和環(huán)境光模型,兩個(gè)模型的相互作用生成了大氣霧霾圖像成像的數(shù)學(xué)模型,模型反推恢復(fù)清晰圖像。最后,進(jìn)行交通標(biāo)志的研究,包括兩個(gè)環(huán)節(jié):一個(gè)是標(biāo)志的定位檢測(cè),另一個(gè)是標(biāo)志的分類(lèi)識(shí)別。在交通標(biāo)志定位檢測(cè)算法上,提出了一個(gè)改進(jìn)的基于顏色信息分割和形狀特征定位的交通標(biāo)志檢測(cè)算法;在標(biāo)志識(shí)別環(huán)節(jié)中采用減小搜索區(qū)域來(lái)縮短時(shí)間消耗的模板匹配算法,完成交通標(biāo)志的識(shí)別工作。文章設(shè)計(jì)的霧霾天氣下交通標(biāo)志檢測(cè)與識(shí)別系統(tǒng),能顯著提高系統(tǒng)在霧霾環(huán)境中標(biāo)志的檢測(cè)識(shí)別準(zhǔn)確率,同時(shí)系統(tǒng)也適用于晴天環(huán)境。
[Abstract]:With the rapid development of intelligent transportation system, road traffic sign recognition, as an important part of intelligent transportation system, has gradually become a hot research topic.However, severe weather such as haze often occurs in North and Northeast China, which makes the existing traffic sign detection and recognition rate seriously reduced, and can not meet the actual needs.In order to improve the detection and recognition rate of traffic signs in haze weather, a set of traffic sign detection and recognition system is designed in this paper.Firstly, the first step of haze weather identification system-haze detection is studied.By studying the differences between a large number of clear images of the same scene and haze images, four characteristics of haze images are obtained: image brightness, image contrast, image identification, and the uniqueness of haze region.A Fisher classifier based on image brightness and image contrast is designed to determine whether the current image is a haze image.Secondly, after the image is certified as a haze image, it needs to be clear and restored.The research of fog image sharpening mainly includes two directions: one is image restoration based on mathematical model and the other is image enhancement based on human vision.In this paper, two classical algorithms are summarized, and the simulation results are summarized based on MATLAB platform.In this paper, the atmospheric light attenuation model and the ambient light model are established by using image restoration technology. The interaction of the two models generates the mathematical model of atmospheric haze image imaging, and the model recovers the clear image.Finally, the study of traffic signs includes two steps: one is the location detection of signs, the other is the classification and recognition of signs.In the traffic sign location detection algorithm, an improved traffic sign detection algorithm based on color information segmentation and shape feature location is proposed.Complete the identification of traffic signs.The traffic sign detection and recognition system under haze weather can improve the detection and recognition accuracy of the system in haze environment, and the system is also suitable for sunny weather.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類(lèi)號(hào)】：U495;TP391.41

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 曾友州;胡瑩;曾偉一;鄭曉霞;;提取數(shù)字圖像邊緣的算法比較[J];成都航空職業(yè)技術(shù)學(xué)院學(xué)報(bào);2009年04期

2 宋建中;;噴霧圖像的自動(dòng)分析[J];光學(xué)機(jī)械;1988年04期

3 涂承媛;曾衍鈞;;醫(yī)學(xué)圖像邊緣快速檢測(cè)的模糊集方法[J];北京工業(yè)大學(xué)學(xué)報(bào);2005年06期

4 常君明;馮，

本文編號(hào)：1731063