光谱学与光谱分析
光譜學與光譜分析
광보학여광보분석
SPECTROSCOPY AND SPECTRAL ANALYSIS
2014年
11期
3095-3099
,共5页
余琦%马晓红%王锐%赵华凤
餘琦%馬曉紅%王銳%趙華鳳
여기%마효홍%왕예%조화봉
激光诱导击穿光谱%主成分分析%分类
激光誘導擊穿光譜%主成分分析%分類
격광유도격천광보%주성분분석%분류
Laser induced breakdow n spectroscopy%Principle component analyses%Classification
将激光诱导击穿光谱（LIBS）技术与主成分分析（PCA ）法相结合用于铝合金分类研究，对Al-Cu系、Al-Si系、Al-Mg-Si系、Al-Zn系四类13种铝合金标准样品进行了分类实验，实验结果证明LIBS-PCA方法可以实现铝合金的快速分类。通过使用LIBS技术激发130个铝合金标准样品得到130个光谱样本，再用主成分分析方法进行降维分析，计算出贡献率最大的三个主成分并计算各光谱的主成分得分绘制在三维空间中，发现光谱样本点按照铝合金的种类发生了明显的汇聚现象，由此确定了三个主成分和铝合金类型区域。用20个不同类型的铝合金进行实验对所得铝合金类型区域的准确性进行验证，发现所得20个光谱样本点全部落在其对应的标准样品类型区域内，在一定程度上证明所得的铝合金标准样品类型区域的正确性，在此基础上可以进行未知类型铝合金的鉴别。实验结果表明基于LIBS光谱的 PCA方法分类精度达到97.14％以上，能够有效的完成不同模式的区分，相比于常用的化学方法，L IBS技术可以原位快速地对待测样品进行检测，样品预处理简单，因此将激光诱导击穿光谱（LIBS）技术与主成分分析（PCA）法相结合用于质量检测和在线工业控制等领域，可以节约大量的时间及成本，提高检测效率。
將激光誘導擊穿光譜（LIBS）技術與主成分分析（PCA ）法相結閤用于鋁閤金分類研究，對Al-Cu繫、Al-Si繫、Al-Mg-Si繫、Al-Zn繫四類13種鋁閤金標準樣品進行瞭分類實驗，實驗結果證明LIBS-PCA方法可以實現鋁閤金的快速分類。通過使用LIBS技術激髮130箇鋁閤金標準樣品得到130箇光譜樣本，再用主成分分析方法進行降維分析，計算齣貢獻率最大的三箇主成分併計算各光譜的主成分得分繪製在三維空間中，髮現光譜樣本點按照鋁閤金的種類髮生瞭明顯的彙聚現象，由此確定瞭三箇主成分和鋁閤金類型區域。用20箇不同類型的鋁閤金進行實驗對所得鋁閤金類型區域的準確性進行驗證，髮現所得20箇光譜樣本點全部落在其對應的標準樣品類型區域內，在一定程度上證明所得的鋁閤金標準樣品類型區域的正確性，在此基礎上可以進行未知類型鋁閤金的鑒彆。實驗結果錶明基于LIBS光譜的 PCA方法分類精度達到97.14％以上，能夠有效的完成不同模式的區分，相比于常用的化學方法，L IBS技術可以原位快速地對待測樣品進行檢測，樣品預處理簡單，因此將激光誘導擊穿光譜（LIBS）技術與主成分分析（PCA）法相結閤用于質量檢測和在線工業控製等領域，可以節約大量的時間及成本，提高檢測效率。
장격광유도격천광보（LIBS）기술여주성분분석（PCA ）법상결합용우려합금분류연구，대Al-Cu계、Al-Si계、Al-Mg-Si계、Al-Zn계사류13충려합금표준양품진행료분류실험，실험결과증명LIBS-PCA방법가이실현려합금적쾌속분류。통과사용LIBS기술격발130개려합금표준양품득도130개광보양본，재용주성분분석방법진행강유분석，계산출공헌솔최대적삼개주성분병계산각광보적주성분득분회제재삼유공간중，발현광보양본점안조려합금적충류발생료명현적회취현상，유차학정료삼개주성분화려합금류형구역。용20개불동류형적려합금진행실험대소득려합금류형구역적준학성진행험증，발현소득20개광보양본점전부락재기대응적표준양품류형구역내，재일정정도상증명소득적려합금표준양품류형구역적정학성，재차기출상가이진행미지류형려합금적감별。실험결과표명기우LIBS광보적 PCA방법분류정도체도97.14％이상，능구유효적완성불동모식적구분，상비우상용적화학방법，L IBS기술가이원위쾌속지대대측양품진행검측，양품예처리간단，인차장격광유도격천광보（LIBS）기술여주성분분석（PCA）법상결합용우질량검측화재선공업공제등영역，가이절약대량적시간급성본，제고검측효솔。
Laser-induced breakdown spectroscopy (LIBS) and the principle component analysis(PCA) were combined to study a-luminum alloy classification in the present article .Classification experiments were done on thirteen different kinds of standard samples of aluminum alloy which belong to 4 different types ,and the results suggested that the LIBS-PCA method can be used to aluminum alloy fast classification .PCA was used to analyze the spectrum data from LIBS experiments ,three principle compo-nents were figured out that contribute the most ,the principle component scores of the spectrums were calculated ,and the scores of the spectrums data in three-dimensional coordinates were plotted .It was found that the spectrum sample points show clear convergence phenomenon according to the type of aluminum alloy they belong to .This result ensured the three principle compo-nents and the preliminary aluminum alloy type zoning .In order to verify its accuracy ,20 different aluminum alloy samples were used to do the same experiments to verify the aluminum alloy type zoning .The experimental result showed that the spectrum sample points all located in their corresponding area of the aluminum alloy type ,and this proved the correctness of the earlier a-luminum alloy standard sample type zoning method .Based on this ,the identification of unknown type of aluminum alloy can be done .All the experimental results showed that the accuracy of principle component analyses method based on laser-induced breakdown spectroscopy is more than 97.14% ,and it can classify the different type effectively .Compared to commonly used chemical methods ,laser-induced breakdown spectroscopy can do the detection of the sample in situ and fast with little sample preparation ,therefore ,using the method of the combination of LIBS and PCA in the areas such as quality testing and on-line in-dustrial controlling can save a lot of time and cost ,and improve the efficiency of detection greatly .
