本文關(guān)鍵詞： 三維測(cè)量 編碼結(jié)構(gòu)光 特征點(diǎn)檢測(cè) 解碼 深度學(xué)習(xí)　出處：《五邑大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：編碼結(jié)構(gòu)光投影三維測(cè)量方法具有無損、成本低、精度高、設(shè)備簡(jiǎn)單的優(yōu)點(diǎn),是目前三維測(cè)量、三維重建應(yīng)用中主要采用的技術(shù)之一。但在編碼的設(shè)計(jì)、特征點(diǎn)匹配、以及解碼的魯棒性方面仍然存在技術(shù)難點(diǎn)。由于測(cè)量對(duì)象表面顏色、紋理和亮度的不確定性,以及光照的影響,動(dòng)態(tài)目標(biāo)的時(shí)間相關(guān)性和空間相關(guān)性較弱,現(xiàn)有的空間編碼結(jié)構(gòu)光編解碼算法容易受被測(cè)量對(duì)象表面顏色、紋理和光照的影響而使其三維測(cè)量魯棒性不高。為此,本文以偽隨機(jī)編碼理論為基礎(chǔ),以二值幾何符號(hào)為編碼符號(hào),設(shè)計(jì)了不易受表面顏色、紋理及光照變化影響、具有抗噪聲能力、編碼容量大、窗口尺寸小的編碼圖案。同時(shí),根據(jù)編碼圖案的特殊結(jié)構(gòu),設(shè)計(jì)了一種多模板特征點(diǎn)檢測(cè)算法,實(shí)驗(yàn)證明此算法抗噪性能強(qiáng),對(duì)具有不同顏色、紋理、光照和曲率的表面仍具有較好的特征點(diǎn)檢測(cè)效果。為最大限度降低畸變圖像特征點(diǎn)檢測(cè)的錯(cuò)誤率,在基于極限約束優(yōu)化的基礎(chǔ)上擴(kuò)展了平面約束和拓?fù)浼s束,提高了特征點(diǎn)檢測(cè)精度。與編碼方案相對(duì)應(yīng),本文設(shè)計(jì)了一種解碼速度快、精度高的解碼方法,此解碼方法利用基于深度學(xué)習(xí)的卷積神經(jīng)網(wǎng)絡(luò)實(shí)現(xiàn)解碼。首先采集大量的不同顏色、材質(zhì)表面的三維測(cè)量目標(biāo)在不同光照條件下的編碼結(jié)構(gòu)光投影的編碼符號(hào)圖像樣本,用于訓(xùn)練卷積神經(jīng)網(wǎng)絡(luò),然后在解碼階段利用訓(xùn)練好的卷積網(wǎng)絡(luò)來對(duì)編碼符號(hào)圖像進(jìn)行分類識(shí)別,實(shí)現(xiàn)對(duì)受目標(biāo)物體表面調(diào)制而形成的畸變圖案的解碼,并進(jìn)而完成三維信息測(cè)量。實(shí)驗(yàn)表明,本文解碼算法對(duì)編碼符號(hào)測(cè)試集識(shí)別率可達(dá)98.07%,對(duì)顏色變化、陰影、紋理結(jié)構(gòu)變化或反射特性不同的動(dòng)態(tài)目標(biāo)、不同尺寸的樣本都具有較好的解碼效果,并且具有一定的抗噪性能。為驗(yàn)證本文設(shè)計(jì)的基于十字的多模板特征點(diǎn)檢測(cè)算法、基于卷積網(wǎng)絡(luò)的解碼算法及優(yōu)化機(jī)制的三維測(cè)量方法性能,選取表面特性不同的物體,分別對(duì)其特征點(diǎn)檢測(cè)效果、點(diǎn)云數(shù)據(jù)、深度圖、重建效果進(jìn)行分析。實(shí)驗(yàn)結(jié)果表明,在結(jié)構(gòu)光三維測(cè)量中,本文提出的編碼模版、特征點(diǎn)檢測(cè)算法、解碼算法表現(xiàn)出了更好的精度和魯棒性,并可以進(jìn)一步應(yīng)用于實(shí)際動(dòng)、靜態(tài)目標(biāo)的高精度三維測(cè)量。
[Abstract]:The coded structured light projection 3D measurement method has the advantages of nondestructive, low cost, high precision and simple equipment. It is one of the main techniques used in the application of 3D measurement and 3D reconstruction. Because of the uncertainty of surface color, texture and brightness, and the influence of illumination, the temporal and spatial correlation of dynamic target is weak. The existing space-coded structured light codec algorithms are easy to be affected by the surface color, texture and illumination of the measured object, which makes the 3D measurement less robust. Therefore, based on the pseudorandom coding theory, Taking binary geometric symbols as coding symbols, a coding pattern with anti-noise ability, large coding capacity and small window size is designed, which is not easily affected by the changes of surface color, texture and illumination. At the same time, according to the special structure of the coding pattern, A multi-template feature point detection algorithm is designed. Experiments show that the algorithm has strong anti-noise performance and has different colors and textures. In order to minimize the error rate of feature point detection in distorted images, plane constraints and topological constraints are extended based on the optimization of limit constraints. The accuracy of feature point detection is improved. Corresponding to the coding scheme, this paper designs a decoding method with high decoding speed and high precision. This decoding method uses convolutional neural network based on deep learning to decode. Firstly, a large number of coded symbol images of different color and material surface 3D measurement objects are coded under different illumination conditions. It is used to train the convolutional neural network, and then use the trained convolution network to classify and recognize the coded symbol image in the decoding stage, so as to decode the distorted pattern formed by the modulation of the target object's surface. The experimental results show that the decoding algorithm can recognize the coded symbol test sets with a rate of 98.07, and the dynamic targets with different color changes, shadows, texture structure changes or reflection characteristics can be obtained. All samples of different sizes have good decoding effect, and have a certain anti-noise performance. In order to verify the multi-template feature point detection algorithm based on cross designed in this paper, Based on the decoding algorithm of convolution network and the performance of 3D measurement method based on optimization mechanism, the feature point detection effect, point cloud data, depth map and reconstruction effect of objects with different surface characteristics are analyzed respectively. The experimental results show that, In structured light 3D measurement, the coding template, feature point detection algorithm and decoding algorithm presented in this paper show better accuracy and robustness, and can be further applied to high precision 3D measurement of moving and static targets.
【學(xué)位授予單位】：五邑大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 王濤;高賢強(qiáng);;基于U3D和kinect1.0月光下舞蹈互動(dòng)游戲的設(shè)計(jì)與實(shí)現(xiàn)[J];西安航空學(xué)院學(xué)報(bào);2017年01期

2 張登攀;田振華;王東升;;三維虛擬動(dòng)態(tài)測(cè)量系統(tǒng)構(gòu)建方法研究[J];計(jì)算機(jī)測(cè)量與控制;2016年05期

3 唐蘇明;張旭;屠大維;;彩色偽隨機(jī)編碼結(jié)構(gòu)光解碼方法研究[J];光電子·激光;2015年03期

4 陸軍;張?chǎng)?高樂;董東來;;基于De Bruijn序列的彩色結(jié)構(gòu)光編解碼方法研究[J];光電子.激光;2014年01期

5 韓成;秦貴和;宮宇;張超;薛耀紅;;基于彩色結(jié)構(gòu)光的三維重構(gòu)方法[J];吉林大學(xué)學(xué)報(bào)(工學(xué)版);2013年05期

6 陸軍;高樂;張?chǎng)?;彩色結(jié)構(gòu)光系統(tǒng)高低強(qiáng)度條紋的顏色聚類方法[J];計(jì)算機(jī)應(yīng)用;2013年08期

7 陸軍;李積江;黃春明;;符號(hào)M陣列結(jié)構(gòu)光的解碼[J];光學(xué)精密工程;2013年04期

8 范靜濤;韓成;張超;李明勛;白寶興;楊華民;;一種新的De Bruijn彩色結(jié)構(gòu)光解碼技術(shù)研究[J];電子學(xué)報(bào);2012年03期

9 李明勛;楊華民;張超;陳展東;韓成;;基于彩色結(jié)構(gòu)光的顏色聚類化方法研究[J];長(zhǎng)春理工大學(xué)學(xué)報(bào)(自然科學(xué)版);2011年02期

10 魏志強(qiáng);高興堂;紀(jì)筱鵬;;基于K均值算法的彩色編碼條紋分色研究[J];計(jì)算機(jī)應(yīng)用;2010年S2期

相關(guān)碩士學(xué)位論文 前6條

1 單夢(mèng)園;基于結(jié)構(gòu)光立體視覺的三維測(cè)量技術(shù)研究[D];哈爾濱工業(yè)大學(xué);2016年

2 李曉東;基于動(dòng)態(tài)編碼點(diǎn)的大型航空構(gòu)件三維型面測(cè)量研究[D];大連理工大學(xué);2016年

3 陳士行;三維動(dòng)態(tài)磨削力測(cè)量平臺(tái)設(shè)計(jì)與研究[D];電子科技大學(xué);2016年

4 劉猛;三維模型動(dòng)態(tài)紋理映射[D];浙江大學(xué);2015年

5 孟影;多視角動(dòng)態(tài)三維重建技術(shù)研究與實(shí)現(xiàn)[D];杭州電子科技大學(xué);2015年

6 賈帥帥;動(dòng)態(tài)三維形貌檢測(cè)關(guān)鍵技術(shù)研究[D];山東大學(xué);2014年

，

本文編號(hào)：1545882