《電子技術(shù)應(yīng)用》
您所在的位置:首頁 > 通信與網(wǎng)絡(luò) > 設(shè)計(jì)應(yīng)用 > 基于特征點(diǎn)提取和PCA的改進(jìn)ICP點(diǎn)云配準(zhǔn)方法
基于特征點(diǎn)提取和PCA的改進(jìn)ICP點(diǎn)云配準(zhǔn)方法
電子技術(shù)應(yīng)用
馬然
廣州南方測(cè)繪科技股份有限公司
摘要: 傳統(tǒng)迭代最近點(diǎn)(Iterative Closest Point, ICP)方法進(jìn)行點(diǎn)云配準(zhǔn)時(shí)存在實(shí)時(shí)性差、易陷入局部極值且配準(zhǔn)精度低等問題。提出一種基于特征點(diǎn)提取、主成分分析(Principal Component Analysis, PCA)粗配準(zhǔn)和ICP精配準(zhǔn)的三步點(diǎn)云配準(zhǔn)方法。首先定義點(diǎn)云數(shù)據(jù)局部密度概念,并自動(dòng)選擇局部密度較大的點(diǎn)作為特征點(diǎn),然后利用PCA對(duì)提取的特征點(diǎn)進(jìn)行分析,根據(jù)PCA主分量方向計(jì)算配準(zhǔn)所需平移和旋轉(zhuǎn)參數(shù)。最后利用ICP對(duì)數(shù)據(jù)進(jìn)行精配準(zhǔn)。試驗(yàn)結(jié)果表明,所提方法相對(duì)于對(duì)比方法的配準(zhǔn)精度提升超過13.4%,實(shí)時(shí)性提升超過38.2%,并且在低信噪比條件下表現(xiàn)出了更高的適應(yīng)性,具有較高的應(yīng)用前景。
中圖分類號(hào):P209 文獻(xiàn)標(biāo)志碼:A DOI: 10.16157/j.issn.0258-7998.245473
中文引用格式: 馬然. 基于特征點(diǎn)提取和PCA的改進(jìn)ICP點(diǎn)云配準(zhǔn)方法[J]. 電子技術(shù)應(yīng)用,2025,51(4):110-115.
英文引用格式: Ma Ran. Improved ICP point cloud registration method based on feature point extraction and PCA[J]. Application of Electronic Technique,2025,51(4):110-115.
Improved ICP point cloud registration method based on feature point extraction and PCA
Ma Ran
Guangzhou Southern Surveying and Mapping Technology Co., Ltd.
Abstract: The traditional Iterative Closest Point (ICP) method for point cloud registration has problems such as poor real-time performance, susceptibility to local extremum, and low registration accuracy. This paper proposes a three-step point cloud registration method based on feature point extraction, Principal Component Analysis (PCA) coarse registration, and ICP fine registration. Firstly, it defines the concept of local density in point cloud data and automatically selects points with higher local density as feature points. Then, it uses PCA to analyze the extracted feature points and calculates the required translation and rotation parameters for registration based on the principal component direction of PCA. Finally, it uses ICP to perform precise data registration. The experimental results show that the proposed method improves registration accuracy by more than 13.4% compared to the comparison methods, improves real-time performance by more than 38.2%, and exhibits higher adaptability under low signal-to-noise ratio conditions, with high application prospects.
Key words : 3D laser;point cloud registration;iteration closest point;local density;principal component analysis

引言

三維激光掃描技術(shù)具有高精度、高分辨率和非接觸等優(yōu)點(diǎn),近年來在醫(yī)療、測(cè)繪、軍事、交通等眾多領(lǐng)域得到廣泛應(yīng)用。由于掃描對(duì)象尺寸大或掃描角度限制等原因,三維激光掃描得到的點(diǎn)云數(shù)據(jù)難以一次性實(shí)現(xiàn)對(duì)掃描對(duì)象的完整描述,通常需要進(jìn)行多次多角度點(diǎn)云數(shù)據(jù)采集,再通過點(diǎn)云配準(zhǔn)算法對(duì)獲得的多次多角度數(shù)據(jù)進(jìn)行配準(zhǔn)才能獲得完整的對(duì)象描述[1-2]。這一過程中,高精度、高實(shí)時(shí)性的點(diǎn)云配準(zhǔn)算法是關(guān)鍵。

迭代最近點(diǎn)(Iterative Closest Point, ICP)算法是Besl等于1992年提出的一種經(jīng)典點(diǎn)云配準(zhǔn)算法[3],也是目前應(yīng)用最為廣泛的一種方法。ICP在多次多角度點(diǎn)云數(shù)據(jù)初始位置相差不大的情況下能夠獲得較高的配準(zhǔn)精度,但是當(dāng)初始位姿差異較大或點(diǎn)云重疊度較低時(shí)算法易陷入局部最優(yōu),實(shí)時(shí)性和配準(zhǔn)精度均會(huì)出現(xiàn)較大程度下降[4-6]。文獻(xiàn)[7]將全局分界支定(Branch-and-bound, BNB)方法引入ICP,提出一種具備全局優(yōu)化能力的BNB-ICP點(diǎn)云配準(zhǔn)算法,能夠提升ICP算法對(duì)初始位置的適應(yīng)性,但是算法運(yùn)算效率較低;文獻(xiàn)[8]提出一種結(jié)合快速點(diǎn)特征直方圖(Fast Point Features Histograms, FPFH)和ICP結(jié)合的點(diǎn)云配準(zhǔn)算法,利用FPFH得到點(diǎn)云特征點(diǎn),并根據(jù)特征點(diǎn)實(shí)現(xiàn)點(diǎn)云粗配準(zhǔn),之后利用ICP進(jìn)行精配準(zhǔn),雖然改善了配準(zhǔn)精度,但是不適合初始位姿較差的情況;文獻(xiàn)[9]將八叉樹算法引入點(diǎn)云配準(zhǔn)領(lǐng)域,利用八叉樹建立不同姿態(tài)點(diǎn)云數(shù)據(jù)之間的拓?fù)潢P(guān)系,進(jìn)而利用ICP完成配準(zhǔn),該算法運(yùn)算效率較高且對(duì)結(jié)構(gòu)簡(jiǎn)單對(duì)象的配準(zhǔn)效果較好,但是不適合結(jié)果復(fù)雜對(duì)象配準(zhǔn);文獻(xiàn)[10]首先計(jì)算點(diǎn)云數(shù)據(jù)的主方向和曲率,并根據(jù)主方向和曲率選擇特征點(diǎn)進(jìn)行粗配準(zhǔn),最后利用ICP進(jìn)行精配準(zhǔn),該方法運(yùn)算效率高,實(shí)時(shí)性好,但是當(dāng)對(duì)象表面結(jié)構(gòu)較為平滑時(shí),即曲率特征不明顯時(shí)該方法的魯棒性較差;文獻(xiàn)[11]將Procrustes正交分解與ICP結(jié)合,利用Procrustes對(duì)點(diǎn)云數(shù)據(jù)進(jìn)行正交分析獲得平移和旋轉(zhuǎn)轉(zhuǎn)換參數(shù),進(jìn)而利用ICP完成點(diǎn)云配準(zhǔn),該方法精度較高且具有較好的魯棒性,但是對(duì)噪聲敏感,不適合低信噪比情況應(yīng)用。

在上述研究的基礎(chǔ)上,本文提出一種基于點(diǎn)云數(shù)據(jù)局部密度提取特征點(diǎn),然后利用PCA對(duì)特征點(diǎn)進(jìn)行投影計(jì)算平移和旋轉(zhuǎn)參數(shù)從而實(shí)現(xiàn)粗配準(zhǔn),最后利用ICP進(jìn)行精配準(zhǔn)的三步配準(zhǔn)方法。利用斯坦福大學(xué)標(biāo)準(zhǔn)數(shù)據(jù)集驗(yàn)證了所提方法的有效性和優(yōu)越性。


本文詳細(xì)內(nèi)容請(qǐng)下載:

http://ihrv.cn/resource/share/2000006405


作者信息:

馬然

(廣州南方測(cè)繪科技股份有限公司, 廣東 廣州 510000)


Magazine.Subscription.jpg

此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載。