基于近紅外光譜的豬肉貯藏時(shí)間判定研究

1.47萬(wàn)字
我自己的畢業(yè)論文，原創(chuàng)的，已經(jīng)通過(guò)校內(nèi)系統(tǒng)檢測(cè)，僅在本站獨(dú)家出售，重復(fù)率低，大家放心下載使用

摘要　本文首先綜述了豬肉新鮮度研究背景及其檢測(cè)現(xiàn)狀，進(jìn)而闡述近紅外光譜技術(shù)的原理、特點(diǎn)，最終綜述了近紅外光譜技術(shù)在肉類檢測(cè)中的重要應(yīng)用以及進(jìn)展, 主要包括影響肉類品質(zhì)的蛋白質(zhì)、脂肪及水分等化學(xué)組成成分分析,嫩度、保水性、肉色及新鮮度等感官指標(biāo)的評(píng)價(jià),并針對(duì)目前發(fā)展趨勢(shì)展望了該技術(shù)的前景。豬肉的貯藏時(shí)間和豬肉的新鮮度緊密相關(guān)。通過(guò)近紅外漫反射光譜技術(shù)獲取豬肉樣本數(shù)據(jù)，利用正交線性判別分析（OLDA）算法進(jìn)行特征提取，本研究將Adaboost算法與OLDA算法結(jié)合，利用Adaboost算法進(jìn)行多輪迭代訓(xùn)練，從而得到多個(gè)OLDA特征提取器，最終采用加權(quán)投票方式重新構(gòu)造了一個(gè)強(qiáng)分類器。這種集成學(xué)習(xí)的特征提取算法不僅分類正確率高，經(jīng)過(guò)3到4輪迭代循環(huán)就能達(dá)到90%的分類正確率，而且多輪迭代的訓(xùn)練時(shí)間仍然接近于傳統(tǒng)的特征提取算法，在效率和性能取得了相對(duì)平衡。
關(guān)鍵詞：近紅外光譜；豬肉；貯藏時(shí)間；adaboost；olda

Study on determination of pork storage time based on near infrared spectroscopy
Abstract The article first reviews the background and research status of pork freshness detection,and then discusses the principle, characteristics of the near infrared spectrum technology,finally summarized the progress of application of near infrared spectroscopy in the meat industry and the research in recent years, mainly including protein, fatand moisture of meat quality effect of chemical component analysis, eva luationmeat sensory quality such as tenderness, water holding capacity, color and freshness index, and according to the development trend of the technology in the future.The storage time of pork and pork freshness of closely related. Gets the porksample data by near infrared diffuse reflectance spectroscopy, using orthogonallinear discriminant analysis (OLDA) algorithm for feature extraction, in this study,Adaboost and OLDA algorithm, multi iteration training by using Adaboost algorithm, so as to obtain multiple OLDA feature extractor, eventually re construct a strong classifier using weighted voting. Feature extraction algorithm for theintegration of learning is not only the correct classification rate is high, after 3 to 4rounds of iterative loop can achieve 90% correct classification rate, and multiple iteration of training time is still close to the traditional feature extraction algorithm,the relative balance between the efficiency and the performance.
Key words: Near Infrared Spectroscopy; Pork; Storage time;adaboost；olda


目錄
第一章　引言 1
1.1　豬肉新鮮度的研究背景 1
1.2　豬肉新鮮度的檢測(cè)現(xiàn)狀 1
1.2.1　感官評(píng)定 1
1.2.2　揮發(fā)性鹽基氮的測(cè)定 2
1.2.3　近紅外光譜法 3
1.2.4　小結(jié) 3
第二章　近紅外光譜技術(shù) 4
2.1　近紅外光譜分析原理 4
2.2　近紅外光譜分析技術(shù)的特點(diǎn) 5
2.3　近紅外光譜分析技術(shù)在肉類檢測(cè)中的應(yīng)用 6
2.3.1　近紅外技術(shù)應(yīng)用于肉類化學(xué)組成分析 6
2.3.2　近紅外技術(shù)應(yīng)用于肉類感官品質(zhì)的評(píng)價(jià) 6
2.3.3　近紅外技術(shù)應(yīng)用于肉類的鑒定和判斷 8
2.4　前景與展望 8
第三章　近紅外光譜特征提取算法 10
3.1　正交線性判別分析（OLDA）原理 10
3.2　Adaboost算法原理 11
3.3　Adaboost+OLDA 算法 12
3.4　Adaboost.M1算法 13
3.5　PCA算法 14
3.6　LDA算法 15
3.7　小結(jié) 16
第四章　實(shí)驗(yàn)與結(jié)果分析 17
4.1　MATLAB簡(jiǎn)介 17
4.2　實(shí)驗(yàn)材料與光譜測(cè)定方法 19
4.3　實(shí)驗(yàn)數(shù)據(jù)采集 20
4.4　OLDA與PCA+LDA對(duì)比實(shí)驗(yàn) 21
4.5　實(shí)驗(yàn)結(jié)合OLDA的特征提取算法對(duì)分類效果的改進(jìn) 24
4.6　Adaboost+OLDA與Adaboost+LDA對(duì)比實(shí)驗(yàn) 26
結(jié) 論 29
致 謝 30
參考文獻(xiàn) 31