Fast and nondestructive gender detection of Bombyx mori chrysalis in the cocoon based on near infrared transmission spectroscopy
-
摘要:目的
利用近红外漫透射光谱快速无损鉴别家蚕Bombyx mori种茧茧壳内蚕蛹的雌雄,以提高育种效率、降低人工成本。
方法以芙9、9芙、湘7和7湘4个蚕品种为研究对象,采集比较了样本在可见和近红外区间的漫透射光谱,建立比较了各品种偏最小二乘判别分析(PLSDA)、后向传播神经网络(BPNN)以及支持向量机分类(SVM)判别模型,通过分类器特性(ROC)曲线研究了各模型的鲁棒性,采用差值法和遗传算法提取了特征波长。
结果芙9、9芙、湘7和7湘品种利用450~900 nm光谱建模的雌雄鉴别准确率分别为95.20%、95.65%、88.80%和87.50%,利用900~1 700 nm 光谱建模的准确率分别为100%、96.00%、92.22%和94.21%;采用PLSDA、BPNN和SVM模型都能够对蚕蛹雌雄做出较好的无损鉴别,3种模型真雌性率分别为95.96%、95.83%和100%,真雄性率分别为98.98%、96.04%和82.18%,准确率分别为97.46%、95.94%和90.86%,进一步通过ROC曲线分析,PLSDA模型效果最优,BPNN模型次之;手动提取20个波段建立PLSDA模型,鉴别真雌性率为93.75%,真雄性率为95.45%,准确率为94.57%。
结论近红外波段900~1 700 nm的漫透射光谱比可见–近红外波段450~950 nm含有更丰富的蚕蛹雌雄分类信息;3种鉴别模型中,PLSDA模型效果最优;提取特征波段后,准确率能达到生产需要。
Abstract:ObjectiveTo identify the gender of Bombyx mori chrysalis in the cocoon by rapid and non-destructive method based on near infrared transmission spectroscopy, improve breeding efficiency and reduce labor cost.
MethodWe used four silkworm varieties including Fu 9, 9 Fu, Xiang 7 and 7 Xiang, and compared their diffuse transmission spectra between 450-950 nm and 900-1700 nm. Partial least squares discrimination analysis (PLSDA), back propagation neural network (BPNN) and support vector machine (SVM) discrimination models were established and compared among different varieties. The robustness of the models was studied through the receiver operating characteristic(ROC) curve. Characteristic wavelengths were extracted by difference method and genetic algorithm.
ResultThe identification accuracy rates for Fu 9, 9 Fu, Xiang 7 and 7 Xiang varieties were 95.20%, 95.65%, 88.80% and 87.50% respectively using 450-950 nm spectra, and were 100%, 96.00%, 92.22% and 94.21% respectively using 900-1 700 nm spectra. Using PLSDA, BPNN and SVM models resulted in good identification of male and female silkworm pupae, the true female rates were 95.96%, 95.83% and 100%, the true male rates were 98.98%, 96.04% and 82.18%, and the accuracy rates were 97.46%, 95.94% and 90.86%, respectively. Based on the analysis of ROC curve, the PLSDA model was the optimal, followed by the BPNN model. Twenty bands were extracted manually as the equipment input, and the true female rate, true male rate and accuracy rate were 93.75%, 95.45% and 94.57% respectively based on the PLSDA model.
ConclusionDiffuse transmission spectra in the near infrared (900-1 700 nm) contains more classification information of male and female pupae compared with the visible-near infrared (450-950 nm). The PLSDA model is the optimal one among three models. After extracting the characristic bands, the accuracy rate can meet the requirements of actual production.
-
Keywords:
- Bombyx mori /
- silkworm chrysalis /
- gender /
- near infrared spectroscopy /
- diffuse transmission /
- band /
- nondestructive detection
-
-
表 1 PLSDA、BPNN和SVM建模方法的效果比较
Table 1 Effect comparison of PLSDA,BPNN and SVM models
模型 样本数 真实值 预测值 真雌性率/% 真雄性率/% 准确率/% AUC1) 雌性 雄性 雌性 雄性 正确 错误 正确 错误 PLSDA 197 96 101 95 4 97 1 95.96 98.98 97.46 0.975 BPNN 197 96 101 92 4 97 4 95.83 96.04 95.94 0.959 SVM 197 96 101 96 18 83 0 100 82.18 90.86 0.910 1) AUC表示分类器特征曲线下的积分面积 表 2 差值法、遗传算法和全波段建模的比较
Table 2 Effect comparison of difference method, genetic algorithm and full-waveband models
方法 样本数 真实值 预测值 真雌性率/% 真雄性率/% 准确率/% AUC1) 雌性 雄性 雌性 雄性 正确 错误 正确 错误 差值法 736 384 352 360 24 336 16 93.75 95.45 94.57 0.972 7 遗传算法 736 384 352 292 92 223 129 76.04 63.35 69.97 0.738 2 全波段 736 384 352 367 17 338 14 95.57 96.02 95.79 0.978 2 1) AUC表示分类器特征曲线下的积分面积 -
[1] 代芬, 吴玲, 叶观燕, 等. 基于近红外漫透射光谱信息的蚕茧雌雄检测[J]. 农业机械学报, 2015, 46(12): 280-283. [2] 封槐松, 李建琴. 我国蚕桑产业发展 “十二五” 回顾与 “十三五” 展望[J]. 中国蚕业, 2016, 37(1): 4-10. [3] 顾国达, 徐俊良. 论我国蚕丝业的多中心起源[J]. 浙江大学学报(人文社会科学版), 2003, 33(3): 42-47. [4] 顾国达, 李建琴. 2012年蚕桑产业发展趋势与建议[J]. 中国蚕业, 2012, 33(1): 1-4. [5] 李建琴, 顾国达. “一带一路” 对中国蚕丝业发展的战略意义[J]. 中国蚕业, 2015, 36(4): 1-7. [6] 潘沈元, 金同铭, 陆国权, 等. 蚕茧近红外反射光谱的模式识别[J]. 生物物理学报, 1995, 11(1): 53-59. [7] 潘沈元. 蚕茧近红外反射光谱的模式识别II. 光谱识别中特征值选取方法的探讨[J]. 生物物理学报, 1998, 14(2): 252-256. [8] 潘沈元, 陶鸣, 孙爱群, 等. 雌雄蚕蛹近红外反射光谱的差异及其模式识别[J]. 昆虫学报, 1996, 39(4): 360-365. [9] MARTEL A, BURGHAMMER M, DAVIES R J, et al. Silk fiber assembly studied by synchrotron radiation SAXS/WAXS and raman spectroscopy[J]. J Am Chem Soc, 2008, 130(50): 17070-17074.
[10] 刘桂花, 张中湖. 拉曼与近红外、中红外光谱在药品掺杂筛选应用中的前景分析[J]. 齐鲁药事, 2012, 31(11): 634-635. [11] LEFEVRE T, PAQUET-MERCIER F, LESAGE S, et al. Study by raman spectromicroscopy of the effect of tensile deformation on the molecular structure of Bombyx mori silk[J]. Vib Spectrosc, 2009, 51(1): 136-141.
[12] 金航峰. 基于光谱和高光谱图像技术的蚕茧品质无损检测研究[D]. 杭州: 浙江大学, 2013. [13] 颜辉, 陈斌, 金尚忠. 基于NIR的主成分结合支持向量机鉴别蚕茧雌雄的研究[J]. 计算机与应用化学, 2008, 25(10): 1261-1264. [14] SHAFFER R E, SMALL G W, ARNOLD M A. Genetic algorithm-based protocol for coupling digital filtering and partial least-squares regression: Application to the near-infrared analysis of glucose in biological matrices[J]. Anal Chem, 1996, 68(15): 2663-2675.
[15] DELWICHE S R, HARELAND G A. Detection of scab-damaged red spring wheat kernels by near-infrared reflectance[J]. Cereal Chem, 2004, 81(5): 643-649.
[16] CHRISTY A A, KASEMSUMRAN S, DU Y, et al. The detection and quantification of adulteration in olive oil by near-infrared spectroscopy and chemometrics[J]. Anal Sci, 2004, 20(6): 935-940.
[17] BERARDO N, PISACANE V, BATTILANI P, et al. Rapid detection of kernel rots and mycotoxins in maize by near-infrared reflectance spectroscopy[J]. J Agr Food Chem, 2005, 53(21): 8128-8134.
[18] BERGHOLT M S, ZHENG W, LIN K, et al. Raman endoscopy for in vivo differentiation between benign and malignant ulcers in the stomach[J]. Analyst, 2010, 135(12): 3162-3168.