共查询到19条相似文献，搜索用时 250 毫秒
1.
基于电子鼻传感器阵列优化的甜玉米种子活力检测 总被引：7，自引：5，他引：2
针对甜玉米种子活力传统检测方法操作繁琐、重复性差等不足,该研究利用电子鼻技术建立甜玉米种子活力快速检测方法。利用电子鼻获取不同活力甜玉米种子的气味信息,再结合主成分分析(PCA,principal component analysis)、线性判别分析(LDA,linear discriminant analysis)、载荷分析(loadings)和支持向量机(SVM,support vector machine)对气味信息进行提取分析,建立甜玉米种子活力的定性定量分析模型。结果显示:PCA和LDA分析均无法区分不同活力的甜玉米种子,而SVM的鉴别效果较好。全传感器阵列数据集SVM分类判别模型训练集和预测集正确率分别为97.10%和96.67%,建模时间为30.75 s,回归预测模型训练集和预测集决定系数R~2分别为0.993和0.913,均方差误差分别为2.23%和8.50%。经Loadings分析将10个传感器阵列优化为6个。优化后传感器阵列数据集SVM分类判别模型训练集和预测集正确率分别为98.55%和96.67%,建模时间为21.81 s,回归预测模型训练集和预测集决定系数R~2分别为0.982和0.984,均方差误差分别为3.80%和3.01%。结果表明:基于SVM的电子鼻技术可以实现对不同活力甜玉米种子的高效判别和预测,将传感器阵列优化为6个,判别和预测效果均有所提升。该研究为电子鼻技术应用于甜玉米种子活力检测提供理论依据。 相似文献
2.
为了研究稻谷中优势霉菌的种类、数量以及变化趋势,本试验以湖南省3个不同地区稻谷为研究对象,通过初步判断各粮仓优势霉菌,结合分子生物学方法对其ITS 序列进行分析,并对稻谷霉菌进行带菌量测定。结果表明,3个粮仓中,金牛仓稻谷中优势菌株分别为根霉、米曲霉、毛霉、黑曲霉、烟曲霉、黄曲霉、白曲霉、桔青霉;金山仓稻谷中优势菌株分别为根霉、米曲霉、毛霉、黑曲霉、黄曲霉、白曲霉、桔青霉;银光仓稻谷中优势菌株分别为根霉、米曲霉、毛霉、黑曲霉、黄曲霉。3个粮仓中根霉数量分布均表现为上层>下层>中层,米曲霉为中层>上层>下层,毛霉、黑曲霉、黄曲霉的分布为上层>中层>下层,同一霉菌在同一粮仓不同粮层间数量存在显著差异(P<0.05);随着储藏时间的延长,米曲霉、黑曲霉、黄曲霉数量逐渐减少,根霉、毛霉数量增加,同一霉菌在不同储藏时间霉菌数量存在差异显著(P<0.05)。本研究结果为稻谷安全储藏与霉菌防控提供了一定的理论依据。 相似文献
3.
基于计算机视觉的稻谷霉变程度检测 总被引：3，自引：1，他引：2
为了实现无损检测稻谷储藏中的霉变,该研究以引起稻谷霉变的5种常见真菌(米曲霉、黑曲霉、构巢曲霉、桔青霉和杂色曲霉)为对象,首先进行真菌培养,制成悬浮液,然后将悬浮液接种到稻谷样品中,对稻谷样品模拟储藏,确定不同霉变程度的稻谷类型,划分为对照组(无霉变)、轻微霉变组和严重霉变组。利用计算机视觉系统对三组稻谷样品进行图像采集和图像处理,提取灰度、颜色和纹理特征,共获取68个图像特征。采用支持向量机(support vector machines,SVM)和偏最小二乘法判别分析(partial least squares discriminant analysis,PLS-DA)构建模型,分别用于无霉变稻谷与霉变稻谷的区分和稻谷霉变类型区分。为了降低模型复杂度和数据冗余,利用连续投影算法(successive projections algorithm,SPA)来消除原始数据变量间的共线性,优选特征值。结果表明:利用所有参数构建的SVM模型能够很好的区分对照组与霉变组,其中建模集和验证集总体区分准确率分别为99.7%和98.4%;SVM模型对于稻谷严重霉变类型的区分效果要优于轻微霉变稻谷,其中对稻谷轻微霉变类型建模集和验证集总体区分的准确率分别为99.3%和92.0%,对稻谷严重霉变类型区分的总体准确率分别为100%和94%,且整体上SVM模型的效果要优于PLS-DA模型。而基于SPA优选特征构建的模型区分结果表明,SVM模型区分效果优于PLS-DA模型,其中,在建模集和验证集中,对无霉变和霉变稻谷总体区分准确率分别为99.8%和99.5%,对稻谷轻微霉变种类区分总体准确率分别为99.8%和90.5%,对稻谷严重霉变种类区分总体准确率分别为100%和95.0%。因此,基于计算机视觉对稻谷霉变检测是可行的,而且SPA优选特征能够较好反映稻谷霉变特征,基于优选特征和SVM模型能够较好地稻谷霉变进行识别和区分,结果较好,可以为实际应用提供技术支持和参考。 相似文献
4.
【目的】利用电子鼻和分光测色仪建立一套快速检测茶树叶片氮含量的无损伤检测方法。【方法】供试样品为茶树顶芽向下第3~4片无损伤叶片。在预实验中优化了气体收集瓶体积、顶空预热温度和顶空时间等参数。采用电子鼻自带Winmuster软件将经过优化后的传感器响应特征值进行主成分分析(principal component analysis,PCA)、线性判别法分析(linear discriminant analysis,LDA)和负荷加载分析(loadings analysis,LA),筛选出灵敏性最好的传感器。同时用分光色差仪对茶树叶片色度值进行测定。样品的测量部位是叶肉区,每组20次重复。色度值主要包括L (表示黑白或者亮暗)、a (表示红绿)、b (表示黄蓝)值。采用Origin 8.0软件对测色仪L、a、b值分别进行一元线性回归分析。利用SPSS 16.0软件采用LSD法进行单因素方差分析(one-way Anova),并进行t检验。对分光测色仪中色差指标进行筛选,以获得相关系数最高的参数。采用凯氏定氮法测定茶叶总氮含量。正式试验第二步是以不同氮含量下的电子鼻和分光测色检测数据为基础,分别建立气味、颜色、气味结合颜色的3种氮含量预测模型,并进行比较分析。【结果】通过预备试验,建立了气体收集器体积为50 mL、顶空预热温度为30℃、顶空时间为30 min的电子鼻检测体系。正式试验第一步确定了以对氮氧化合物灵敏(S2),对甲烷灵敏(S6),对无机硫化物灵敏(S7),对醇类、醛类、酮类物质灵敏(S8),对有机硫化物灵敏(S9)的传感器为主要传感器。根据L、a、b表色系统,b值与叶片缺氮程度呈线性相关。正式试验第二步利用气味、颜色、气味结合颜色建立的3个氮含量预测模型都具有可行性,其中气味结合颜色建立的预测模型准确率最高,达到90%。【结论】用气味结合颜色的预测模型预测茶树叶片氮含量准确度较高,可在实际工作中进行运用。 相似文献
5.
为解析种植至储藏期花生受黄曲霉侵染及黄曲霉毒素B1(AFB_1)污染的规律,揭示花生AFB_1污染的源头及主要影响因素,选取种植湛红2号和湛油75的花生田,采集种植期花生果和土壤及储藏1~4个月的花生果,分析花生和土壤真菌菌相,并采用高效液相色谱法测定花生AFB_1含量。结果表明,种植期黄曲霉侵染花生果主要发生在成熟期,但黄曲霉污染率均在8%以下;湛油75花生田土壤黄曲霉菌落数显著低于湛红2号,但花生果黄曲霉污染率显著高于湛红2号,表明湛红2号具有一定的黄曲霉抗性;湛油75和湛红2号分别在110 d和120 d检测到AFB_1,含量分别为3.37μg·kg~(-1)和2.08μg·kg~(-1),表明花生黄曲霉毒素含量与污染率呈正相关。储藏期花生果中未检测到黄曲霉和AFB_1,这主要是由于花生晾晒后水活度(aw)降低至0.70以下,不适合黄曲霉生长繁殖和毒素生物合成。综上,黄曲霉在荚果成熟期开始侵染花生果导致产生AFB_1,而储藏期保持较低的aw可有效预防黄曲霉及AFB_1污染。本研究结果为制定种植至储藏期花生黄曲霉毒素全程防控措施提供了理论依据。 相似文献
6.
用电子鼻区分霉变燕麦及其传感器阵列优化 总被引：5，自引：4，他引：1
应用电子鼻对燕麦（Avena sativa L）霉变程度进行区分,为了提高区分准确度,对电子鼻传感器阵列进行了优化的研究。每天随机选择10个燕麦样品进行电子鼻检测,试验连续进行5 d,将检测数据耦合入非线性双稳态随机共振系统,以外部Gaussian白噪声激励系统产生共振,选择输出信噪比特征值进行主成分分析,初期试验主成分1和主成分2贡献率之和为96.43%,且相同霉变程度样品离散度较大,不同霉变程度样品之间距离较近。为了提高电子鼻对霉变燕麦样品区分效果,进行了电子鼻传感器负荷加载分析,优化选择了传感器阵列,优化后主成分1和主成分2贡献率之和为99.31%,相同霉变程度燕麦样品的聚合度更高,使不同霉变程度燕麦样品之间的区分更加明显,为进一步的定量化检测奠定了基础。 相似文献
7.
常规稻与杂交稻谷的仿生电子鼻分类识别 总被引：5，自引：5，他引：0
气味是进行稻谷品种及其品质识别的重要方法之一,作为一种基于仿生嗅觉的机器检测方法,仿生电子鼻在水稻品种的分类识别中具有较好的应用前景。常规稻与杂交稻在食味品质等方面存在一定的差异,为了解应用电子鼻进行常规稻谷与杂交稻谷识别的可行性,采用PEN3电子鼻对同季同地域收获的3种常规稻(中香1号、湘晚13、瑶平香)和3种杂交稻(伍丰优T025、品36、优优122)稻谷样品的气味信息进行了采集和分析。首先通过过载分析(Loadings)法分析了电子鼻检测稻谷气体挥发物时的各传感器贡献率,分别针对基于特征值的提取和稻谷气味检测对电子鼻传感器阵列中的传感器进行了优选,阐明了稻谷气体挥发物检测中应以对硫化物、氮氧化合物、芳香成分和有机硫化物敏感的传感器为主。随后,分别采用主成分分析法(principal component analysis,PCA)、线性判别法(linear discriminant analysis,LDA)和BP神经网络对6种不同稻谷之间、常规稻与杂交稻之间的分类识别进行了研究。结果表明,PCA分析法与LDA分析法在对6种不同稻谷之间的分类以及常规稻与杂交稻之间的分类中均未取得理想的效果,存在部分样本数据点重叠或样本数据点较近的情况,在实际应用中易发生混淆;而BP神经网络在对6种不同稻谷之间的分类中对测试集的识别正确率分别达到了90%,在常规稻与杂交稻之间的分类识别中对测试集的识别正确率达到了96.7%。上述试验验证了电子鼻用于常规稻与杂交稻稻谷分类识别的有效性,为常规稻与杂交稻的快速、无损分类识别提供了一种新的方法。 相似文献
8.
电子鼻快速检测谷物霉变的研究 总被引：26，自引：11，他引：26
针对目前我国在谷物的霉变与否的检测上还有一定的滞后性，研制出一套能快速检测谷物是否霉变的电子鼻装置，该装置能快速、准确地分析所测谷物散发的气味，从而判定所测谷物是否霉变。该电子鼻主要由一组厚膜金属氧化锡气体传感器阵列和RBF神经网络组成。用所研制的电子鼻对小麦、水稻、玉米3种谷物进行检测。整个实验过程如下：首先从每个传感器的反应曲线中提取4个特征值，并对所有特征值进行归一化处理，然后用常规的主成分分析和径向基函数(RBF)神经网络对它们进行分析。实验过程中发现，从主成分分析的结果发现很难将霉变谷物与正常谷 相似文献
9.
电子鼻检测鸡蛋货架期新鲜度变化 总被引：19，自引：8，他引：11
该文旨在通过气味检测鸡蛋的新鲜度。利用德国AIRSENSE公司PEN3型电子鼻对鸡蛋在20℃,70%相对湿度条件下罗曼鸡蛋货架期的气味进行了无损检测。通过测定哈夫单位,建立了不同货架期气味与鸡蛋哈夫单位等级的对应关系。首先,分析并对比了第0天与第36天的完整鸡蛋与蛋液所产生气体的变化情况,确定氨氧化物、烷烃和醇类等是鸡蛋贮藏中产生的恶化气体。其次,结合电子鼻,利用主成分分析、线性判别等多元统计方法进行数据分析,对不同货架期、不同等级的鸡蛋进行归类区分,发现线性判别（LDA）效果优于主成分分析法（PCA）。结合载荷分析,确认了检测鸡蛋新鲜度的主要传感器S1、S2、S3、S5、S6、S8。初步证明了气体传感器和模式识别方法在电子鼻区分鸡蛋货架期新鲜度的可行性,为建立利用气体传感器监控鸡蛋新鲜度的方法提供实验基础和理论依据。 相似文献
10.
基于神经网络及电子鼻的虾夷扇贝鲜活品质评价及传感器的筛选 总被引：4，自引：4，他引：0
追踪检测虾夷扇贝品质变化过程中的存活指标,生理指标以及电子鼻气味图谱的变化,建立保活流通过程中不同等级的活品虾夷扇贝电子鼻气味指纹图谱,购买市场上不同状态的活品虾夷扇贝,分别通过学习向量量化(learning vector quantization,LVQ)、概率(probabilistic neural networks,PNN)、支持向量机(support vector machine,SVM)神经网络对测试样品快速模式分类,最后通过对电子鼻传感器的筛选探索便携式快速品质鉴别设备的可能性。研究结果表明,24 h的极端胁迫环境放置较为完整的模拟了虾夷扇贝在保活流通过程中状态变差的过程;将电子鼻数据主成分分析、聚类分析结果与存活指标(开口率、缩边率以及死亡率)和生理指标(超氧化物歧化酶活性、耗氧率以及海水浊度)相结合可以把品质变化过程中的虾夷扇贝分成5个等级,并分别得到每个等级的扇贝气味指纹图谱;3种神经网络均可以对测试样品等级进行快速测定,其中支持向量机(SVM)神经网络兼具精确和快速的特点,测试样本T全部预测为等级4,测试样本N全部预测为等级3,从交叉验证到仿真预测所用时间仅为7.652 s;筛选得到的8个电子鼻传感器也可以对不同等级鲜活虾夷扇贝气味特征进行有效区分。 相似文献
11.
Aflatoxins in domestic and imported foods and feeds 总被引：4，自引：0，他引：4
G E Wood 《Journal of the Association of Official Analytical Chemists》1989,72(4):543-548
Aflatoxins, metabolic products of the molds Aspergillus flavus and A. parasiticus, may occur in foods and feeds. These toxins cannot be entirely avoided or eliminated from foods or feeds by current agronomic and manufacturing processes and are considered unavoidable contaminants. To limit aflatoxin exposure, the U.S. Food and Drug Administration (FDA) has set action levels for these toxins in foods and feeds involved in interstate commerce. FDA continually monitors food and feed industries through compliance programs. This report summarizes data generated from compliance programs on aflatoxins for the fiscal year 1986. Commodities sampled included peanuts and peanut products, corn and corn products, tree nuts, cottonseed, milk, spices, manufactured products, and miscellaneous foods and feeds. Correlations were highest between aflatoxin contamination and geographical areas for corn/corn products and cottonseed/cottonseed meal. Higher incidences of aflatoxin contamination in corn and corn products designated for human consumption were observed in samples collected in the southeastern states (32 and 28%, respectively). A higher incidence of contamination was observed in corn designated for animal feed from Arkansas-Texas (74%) than from the southeastern states (47%). Only 3% of feed corn from corn belt states contained detectable aflatoxins. All aflatoxin-contaminated cottonseed was collected in the Arizona-California area; 80% of cottonseed meal analyzed from this area also contained detectable levels of aflatoxins.(ABSTRACT TRUNCATED AT 250 WORDS) 相似文献
12.
M F Dutton K Westlake 《Journal of the Association of Official Analytical Chemists》1985,68(5):839-842
During the period 1982-1983, just under 800 samples of agricultural commodities, comprising cereals, compound feeds, hay, and silage, were examined for molds and mycotoxins. Aflatoxin B1 showed the highest incidence rate; it occurred in over 27% of all samples analyzed, the highest levels being found in peanut meal at 1500 ppb. Other mycotoxins detected were patulin and a number of trichothecene toxins at incidence rates in all commodities of 5.6 and 3.1%, respectively. The commodities at highest risk were oil seeds, excluding soya bean; the latter was found to be fairly free from contamination with mycotoxins. The most prevalent fungi were Aspergillus flavus and parasiticus, which were found in over 22% of all samples, whereas Penicillium spp. showed the lowest incidence of genera, specifically identified in 8.3% of all samples examined. This latter finding explains in part the low incidence of Penicillium mycotoxins. 相似文献
13.
Liquid chromatographic method for determination of aflatoxins B1, B2, G1, and G2 in corn and peanut products: collaborative study 总被引：1，自引：0，他引：1
D L Park S Nesheim M W Trucksess M E Stack R F Newell 《Journal of the Association of Official Analytical Chemists》1990,73(2):260-266
A collaborative study of a liquid chromatographic method for the determination of aflatoxins B1, B2, G1, and G2 was conducted in laboratories located in the United States, Canada, South Africa, and Switzerland. Twenty-one artificially contaminated raw peanuts, peanut butter, and corn samples containing varying amounts of aflatoxins B1, B2, G1, and G2 were distributed to participating laboratories. The test portion was extracted with methanol-0.1N HCl (4 + 1), filtered, defatted with hexane, and then partitioned with methylene chloride. The concentrated extract was passed through a silica gel column. Aflatoxins B1 and G1 were derivatized with trifluoroacetic acid, and the individual aflatoxins were determined by reverse-phase liquid chromatography with fluorescence detection. Statistical analysis of the data was performed to determine or confirm outliers, and to compute repeatability and reproducibility of the method. For corn, relative standard deviations for repeatability (RSDr) for aflatoxin B1 ranged from 27.2 to 8.3% for contamination levels from 5 through 50 ng/g. For raw peanuts and peanut butter, RSDr values for aflatoxin B1 were 35.0 to 41.2% and 11.2 to 19.1%, respectively, for contamination levels from 5 through 25 ng/g. RSDr values for aflatoxins B2, G1, and G2 were similar. Relative standard deviations for reproducibility (RSDr) for aflatoxin B1 ranged from 15.8 to 38.4%, 24.4 to 33.4%, and 43.9 to 54.0% for corn, peanut butter, and raw peanuts, respectively. The method has been adopted official first action for the determination of aflatoxins B1, B2, G1, and G2 in peanut butter and corn at concentrations greater than or equal to 13 ng total aflatoxins/g. 相似文献
14.
Mellon JE Zelaya CA Dowd MK Beltz SB Klich MA 《Journal of agricultural and food chemistry》2012,60(10):2740-2745
Racemic gossypol and its related derivatives gossypolone and apogossypolone demonstrated significant growth inhibition against a diverse collection of filamentous fungi that included Aspergillus flavus, Aspergillus parasiticus, Aspergillus alliaceus, Aspergillus fumigatus, Fusarium graminearum, Fusarium moniliforme, Penicillium chrysogenum, Penicillium corylophilum, and Stachybotrys atra. The compounds were tested in a Czapek agar medium at a concentration of 100 μg/mL. Racemic gossypol and apogossypolone inhibited growth by up to 95%, whereas gossypolone effected 100% growth inhibition in all fungal isolates tested except A. flavus. Growth inhibition was variable during the observed time period for all tested fungi capable of growth in these treatment conditions. Gossypolone demonstrated significant aflatoxin biosynthesis inhibition in A. flavus AF13 (B(1), 76% inhibition). Apogossypolone was the most potent aflatoxin inhibitor, showing greater than 90% inhibition against A. flavus and greater than 65% inhibition against A. parasiticus (B(1), 67%; G(1), 68%). Gossypol was an ineffectual inhibitor of aflatoxin biosynthesis in both A. flavus and A. parasiticus. Both gossypol and apogossypolone demonstrated significant inhibition of ochratoxin A production (47%; 91%, respectively) in cultures of A. alliaceus. 相似文献
15.
16.
用于污染黄曲霉毒素花生分选的荧光信号研究 总被引：2，自引：2，他引：0
为在加工前将黄曲霉毒素超限的带衣花生米从原料中剔除,参照已有的色选系统,提出一种依据黄曲霉毒素含量超限带衣花生米的专属荧光信号进行逐粒分选的技术构想。采用Cary Eclipse荧光分光光度计测定100粒外观具有代表性的带衣花生米表面的紫外-荧光规律,通过与免疫亲和层析净化荧光光度法(GB/T18979-2003)检测结果对比,判定了黄曲霉毒素超限带衣花生米的荧光光谱特征;通过绘制450/490、460/490荧光强度比值的箱线图,评估了表面荧光法判断黄曲霉毒素超限带衣花生米的准确率;在搭建的荧光成像系统上,对黄曲霉毒素超限带衣花生米进行了荧光成像。检测发现,在365 nm波长激发下,黄曲霉毒素超限带衣花生米在420~460 nm处有荧光峰;以450/490荧光强度比值为依据剔除超限值带衣花生米的判断准确率为81%;a.u.40的带衣花生米可在图像中呈现亮蓝荧光光斑。表明表面荧光信号可作为带衣花生米在线、无损、逐粒分选的专属光学信号,用于黄曲霉毒素超限带衣花生米的剔除。 相似文献
17.
为了解我国花生土壤黄曲霉分布及产毒特征与产后花生黄曲霉毒素污染相关性,本试验从我国4个典型花生产区黄河流域产区(河北保定)、西北产区(新疆吐鲁番)、长江流域产区(湖北黄冈、四川南充)和东南沿海产区(广东湛江)采集花生土壤样品124份。通过对我国不同产区花生土壤中黄曲霉菌的分布及产毒特征研究,评估我国不同产区花生黄曲霉毒素污染风险。结果表明,湖北黄冈和四川南充土壤中黄曲霉检出率、带菌量和黄曲霉毒素B1(AFB1)量均高于其他地区,产后花生受AFB1污染风险最高,其次为广东湛江和新疆吐鲁番,河北保定花生受AFB1污染风险最低。从上述4个产区收集64份花生样品开展产后花生AFB1污染调查,发现湖北黄冈花生AFB1污染情况最严重,检出率为57%,超标率为10%,其次为四川南充和广东湛江,新疆吐鲁番和河北保定花生受AFB1污染较轻。风险评估结果与实际检测结果一致,表明花生土壤中黄曲霉菌数量、产毒菌比例及产毒能力与产后花生AFB1污染呈正相关性。本研究结果为花生黄曲霉毒素污染预警及综合防控提供了理论依据和数据支撑。 相似文献
18.
Gordon SH Jones RW McClelland JF Wicklow DT Greene RV 《Journal of agricultural and food chemistry》1999,47(12):5267-5272
An urgent need for rapid sensors to detect contamination of food grains by toxigenic fungi such as Aspergillus flavus prompted research and development of Fourier transform infrared photoacoustic spectroscopy (FTIR-PAS) as a highly sensitive probe for fungi growing on the surfaces of individual corn kernels. However, the photoacoustic technique has limited potential for screening bulk corn because currently available photoacoustic detectors can accommodate only a single intact kernel at a time. Transient infrared spectroscopy (TIRS), on the other hand, is a promising new technique that can acquire analytically useful infrared spectra from a moving mass of solid materials. Therefore, the potential of TIRS for on-line, noncontact detection of A. flavus contamination in a moving bed of corn kernels was explored. Early test results based on visual inspection of TIRS spectral differences predict an 85% or 95% success rate in distinguishing healthy corn from grain infected with A. flavus. Four unique infrared spectral features which identified infected corn in FTIR-PAS were also found to be diagnostic in TIRS. Although the technology is still in its infancy, the preliminary results indicate that TIRS is a potentially effective screening method for bulk quantities of corn grain. 相似文献
19.
用于山核桃陈化时间检测的电子鼻传感器阵列优化 总被引：7，自引：7，他引：0
为更好地进行山核桃陈化时间检测,论文拟通过传感器阵列优化来有效提高电子鼻对其区分预测能力。该文依据响应曲线保留响应明显的传感器,并在提取传感器特征值构成初始特征矩阵的基础上,结合均值分析、变异系数分析、聚类分析、相关性分析和多重共线性分析进行逐步优化以获取最终优化传感器阵列。对优化前后的数据采用主成分分析法(principal component analysis,PCA)和偏最小二乘回归(partial least squares regression,PLSR)进行样品区分和预测能力的对比。结果表明:通过优化,经不同人工陈化时间(0、5、10、15d)处理的山核桃能有效区分开,且在PCA得分图中更为聚集;优化后的陈化时间回归模型(R2=0.933 4)较优化前(R2=0.888 7)具有更好的预测能力。说明所给出的阵列优化方法有效可行,为电子鼻针对性检测提供了一种思路。 相似文献