于曉洋+潘宗瑋+孫曉明+劉野
摘 要:針對(duì)陶瓷、金屬等光滑表面在結(jié)構(gòu)光三維測(cè)量時(shí),由于表面局部的強(qiáng)反射特性使得物體失真產(chǎn)生較大測(cè)量誤差這一問(wèn)題。依據(jù)反射分量分離理論,設(shè)計(jì)了一種反射分量分離結(jié)合像素填補(bǔ)的改進(jìn)算法。首先通過(guò)像素參數(shù)對(duì)比方法找到圖像中的高光像素點(diǎn),然后對(duì)其進(jìn)行反射分量分離處理,為了避免一些高光強(qiáng)度過(guò)大的像素點(diǎn)在處理之后變?yōu)楹邳c(diǎn),本文采用像素填補(bǔ)方法對(duì)其進(jìn)行處理。實(shí)驗(yàn)結(jié)果表明,采用此種算法對(duì)1920×1080分辨率的圖像進(jìn)行處理,其高光像素的數(shù)目降低了8~11倍,實(shí)現(xiàn)了三維測(cè)量領(lǐng)域中強(qiáng)反射表面高光的有效去除。
關(guān)鍵詞:三維測(cè)量;結(jié)構(gòu)光;反射分量分離;像素填補(bǔ)
DOI:10.15938/j.jhust.2017.03.013
中圖分類(lèi)號(hào): TP391.41
文獻(xiàn)標(biāo)志碼: A
文章編號(hào): 1007-2683(2017)03-0073-07
Abstract:In structured light 3D measurement field, when the object is smooth surface, it can form a highlight area due to the specular reflection of the surface, and the distortion of the object will make a large measurement error. In order to solve this problem, a new algorithm is designed which is based on the theory of reflection component separation and pixel filling. Firstly, the specular pixels in the image are found by comparing the pixel parameters, Then the reflection component is separated and processed. Finally, in order to avoid some of the highlight are processed into black pixels the pixel filling method is adopted to deal with them. Experimental results show that when we deal with the image of 1920*1080 by using this algorithm, the number of specular pixels is reduced by 8-11 times. This method can effectively remove the specular pixels in the field of 3D measurement.
Keywords:structured light; 3D measurement; reflection component separation; pixel filling
4 結(jié) 語(yǔ)
本文設(shè)計(jì)了一種反射分量分離結(jié)合像素填補(bǔ)的改進(jìn)算法,并進(jìn)行了試驗(yàn)系統(tǒng)的搭建,闡述了該方法去除圖像中高光信息的原理。 與傳統(tǒng)的方法相比,該方法完全基于顏色信息,對(duì)被測(cè)物要求較寬松,測(cè)量范圍較廣。實(shí)驗(yàn)結(jié)果表明,對(duì)1920×1080的圖像進(jìn)行處理,其高光像素的數(shù)目降低了8~11倍,能夠有效地去除圖像中的高光像素,驗(yàn)證了該算法的有效性。但本文所提算法對(duì)于高光區(qū)域較大的圖像處理效果不明顯,會(huì)出現(xiàn)失真現(xiàn)象。因此,如何處理高光區(qū)域較大的圖像是今后研究的重點(diǎn)。
參 考 文 獻(xiàn):
[1] 于曉洋,單鸝娜,曹沈楠.結(jié)構(gòu)光時(shí)間編碼技術(shù)進(jìn)展[J].哈爾濱理工大學(xué)學(xué)報(bào),2010, 15(1)1:98-102.
[2] 王旭穎.強(qiáng)反射表面編碼光測(cè)量的高光修復(fù)與抑制方法研究[D].哈爾濱:哈爾濱理工大學(xué),2013:16-19.
[3] 張?chǎng)?,?xí)俊通.基于隨機(jī)編碼結(jié)構(gòu)光的雙目立體三維測(cè)量系統(tǒng)[J].機(jī)電一體化,2013(3):43-54.
[4] 趙鑫,潘晉孝,劉賓.基于線結(jié)構(gòu)光的鋼化玻璃表面形貌的測(cè)量[J].信息與電腦,2013(8):17-18.
[5] 吳慶華.基于線結(jié)構(gòu)光掃描的三維表面缺陷在線檢測(cè)的理論與應(yīng)用研究[D].武漢:華中科技大學(xué),2013:24-30.
[6] 韓書(shū)娜,幺宏偉,范劍英. 指紋圖像分割與增強(qiáng)算法[J].哈爾濱理工大學(xué)學(xué)報(bào). 2011,6(1):98-101.
[7] 楊永敏.強(qiáng)反射表面缺陷圖像預(yù)處理[J].光學(xué)精密工程,2010(18):2288-2296.
[8] 柴玉亭,王昭,高建民.基于頻域?yàn)V波的高光去除方法[J].激光與光電子學(xué)進(jìn)展. 2013, 5:135-143.
[9] KLINKER G J, SHAFER S A, KANADE T. The Measurement of Highlights in Color Images[J]. International Journal of Computer Vision, 1988(2):7-32.
[10]JAKLIC A,SOLINA F. Separating Diffuse Andspecular Component of Image Irradiance by Translating a Camera[C]. CAIP, 1993:428-435.
[11]LIN S,SANG L S. Estimation of Diffuse and Specular Appearance[C]//IEEE International Conference on Computer Vision, 2002,2:855.
[12]黃梅,吳志勇,梁敏華.暗背景下低灰度圖像的實(shí)時(shí)增強(qiáng)[J].液晶與顯示, 2011(3):374-378
[13]王洋.基于中國(guó)剩余定理的模擬編碼結(jié)構(gòu)光三維測(cè)量方法研究[D].哈爾濱:哈爾濱理工大學(xué). 2014:53-69.
[14]MALIICK S P, ZICKLER T E, KRIEGMAN D J,et al. Beyond Lambert: Reconstructing Specular Surfaces Using Color[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2005.
[15]陽(yáng)鵬程.基于雙目結(jié)構(gòu)光的三維測(cè)量技術(shù)研究[D].杭州:浙江大學(xué),2014:32-46.
[16]詹永鑫,鄭學(xué)仁,陳建賓.基于亮度分量的快速圖像增強(qiáng)方法[J].液晶與顯示,2010(2):273-277.
[17]李莎莎,鄧彩霞,侯杰.基于圖像融合技術(shù)的階梯型邊界檢測(cè)[J].哈爾濱理工大學(xué)學(xué)報(bào),2009,14(2):75-77.
[18]韓福海.基于結(jié)構(gòu)光的雙目視覺(jué)三維表面重建[D].北京:北京理工大學(xué),2015: 27-40.
[19]車(chē)建強(qiáng),戴寧,崔海華.面向彩色物體的雙目三維測(cè)量編碼新方法[J].光學(xué)學(xué)報(bào),2014(11):65-68.
[20]孟凡辰,陳德運(yùn),王莉莉.多灰度級(jí)格雷碼三維測(cè)量方法[J].哈爾濱理工大學(xué)學(xué)報(bào).2015,20(5):97-102.
[21]ROBBY T, TAN, KATSUSHI L. Separating Reflection Components of Textured Surfaces Using a Single Image[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005(2): 870-877.
[22]SHAFER S.Using Color to Separate Reflection Components[J]. Color Research and Applications, 1985(10):210-218.
(編輯:溫澤宇)