發(fā)布時(shí)間：2019-01-10 09:53

【摘要】：室內(nèi)環(huán)境下的場(chǎng)景理解是室內(nèi)移動(dòng)機(jī)器人必須具備的能力之一,隨著全球服務(wù)機(jī)器人行業(yè)的興起,半結(jié)構(gòu)化環(huán)境下的室內(nèi)場(chǎng)景理解成為計(jì)算機(jī)視覺(jué)領(lǐng)域的研究熱點(diǎn),也是一個(gè)難點(diǎn),其主要體現(xiàn)在室內(nèi)環(huán)境的復(fù)雜性,識(shí)別算法的魯棒性以及實(shí)時(shí)性上。室內(nèi)場(chǎng)景理解包括室內(nèi)環(huán)境下的目標(biāo)物體檢測(cè),機(jī)器人所處環(huán)境估計(jì),室內(nèi)障礙物規(guī)避,人的檢測(cè)和身份識(shí)別等。圍繞上文所提出的問(wèn)題,本文以室內(nèi)行人和物體檢測(cè)為研究?jī)?nèi)容,主要的研究和工作內(nèi)容如下:1)本文詳細(xì)分析了卷積神經(jīng)網(wǎng)絡(luò)的特征提取和分類方法,并將該方法進(jìn)行物體識(shí)別效果與SIFT特征提取加FLANN匹配方法的物體識(shí)別效果作對(duì)比,得出在目標(biāo)物體的不同觀察角度與目標(biāo)物體發(fā)生形變的情況下,卷積神經(jīng)網(wǎng)絡(luò)物體識(shí)別效果明顯優(yōu)于SIFT特征提取加FLANN匹配方法識(shí)別效果的結(jié)論。2)針對(duì)傳統(tǒng)場(chǎng)景識(shí)別底層特征語(yǔ)義信息表達(dá)能力的不足,結(jié)合卷積神經(jīng)網(wǎng)絡(luò),本文提出一種基于物體檢測(cè)的室內(nèi)場(chǎng)景識(shí)別方法。該方法首先采用卷積神經(jīng)網(wǎng)絡(luò)對(duì)場(chǎng)景中的目標(biāo)進(jìn)行特征提取和分類,然后基于概率模型以檢測(cè)到的目標(biāo)作為中間橋梁去推斷當(dāng)前所處的場(chǎng)景。與基于計(jì)算機(jī)視覺(jué)底層特征的場(chǎng)景識(shí)別方法相比,該方法更接近于人類對(duì)場(chǎng)景的認(rèn)知思維。本文運(yùn)用該方法對(duì)場(chǎng)景的五種室內(nèi)場(chǎng)景進(jìn)行場(chǎng)景識(shí)別分類,取得不錯(cuò)效果。3)為了測(cè)試機(jī)器人在室內(nèi)環(huán)境下對(duì)行人檢測(cè)效果和響應(yīng),本文在PR2機(jī)器人平臺(tái)下基于ROS系統(tǒng)(Robot Operating System),采用Haar-Like特征與Ada Boost分類器實(shí)現(xiàn)人臉檢測(cè),并用EigenFace進(jìn)行身份識(shí)別,同時(shí)用HOG(Histogram of Oriented Gradient)特征與SVM(Support Vector Machine)分類器進(jìn)行人體檢測(cè),并實(shí)現(xiàn)機(jī)器人對(duì)行人的自主跟隨。
[Abstract]:Scene understanding in indoor environment is one of the necessary capabilities of indoor mobile robot. With the rise of global service robot industry, indoor scene understanding in semi-structured environment has become a research hotspot in the field of computer vision. It is also a difficult point, which is mainly reflected in the complexity of indoor environment, the robustness of recognition algorithm and the real-time performance. Indoor scene understanding includes object detection in indoor environment, robot environment estimation, indoor obstacle avoidance, human detection and identity recognition. The main contents of this paper are as follows: 1) the feature extraction and classification methods of convolution neural network are analyzed in detail. The object recognition effect of this method is compared with that of SIFT feature extraction and FLANN matching method. The result of object recognition based on convolution neural network is obviously better than that of SIFT feature extraction and FLANN matching. 2) aiming at the deficiency of semantic information expression of traditional scene recognition underlying features, we combine convolutional neural network with convolution neural network. This paper presents a method of indoor scene recognition based on object detection. Firstly, the convolution neural network is used to extract and classify the features of the targets in the scene, and then based on the probabilistic model, the detected target is used as the intermediate bridge to infer the current scene. Compared with the scene recognition method based on the underlying features of computer vision, this method is more similar to the cognitive thinking of the scene. In this paper, we use this method to classify the scene of five indoor scenes, and get good results. 3) in order to test the robot in the indoor environment to detect the effect and response to pedestrian, In this paper, based on ROS system (Robot Operating System), Haar-Like feature and Ada Boost classifier are used to realize face detection based on PR2 robot platform, and EigenFace is used to identify human face. At the same time, the HOG (Histogram of Oriented Gradient) feature and SVM (Support Vector Machine) classifier are used to detect the human body, and the robot can follow the pedestrian autonomously.
【學(xué)位授予單位】：杭州電子科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41;TP242

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 鄧中亮;余彥培;袁協(xié);萬(wàn)能;楊磊;;室內(nèi)定位現(xiàn)狀與發(fā)展趨勢(shì)研究(英文)[J];中國(guó)通信;2013年03期

，

本文編號(hào)：2406198