基于人工嗅觉的粮食霉变识别方法的研究

发霉的粮食含有对人、畜有危害的霉菌毒素，为了对粮食是否发霉提供一种简单客观的判别方法，研制了一套人工嗅觉系统。该系统主要由气体传感器阵列、气路系统、信号调理电路、数据采集系统及模式识别软件组成。为了提高识别的准确率，利用采集数据中最大响应点及其左右各相隔一定时间的两个点作为3个特征值，并采用3层优化BP神经网络对样本特征值进行训练。经测试，训练样品的回判准确率和测试样品的准确率均为100％，说明该人工嗅觉系统是准确、有效的。     

电子鼻与电子舌在食品检测中的应用研究进展

随着嗅觉与味觉传感器技术的发展，电子鼻与电子舌技术在食品检测中得到了不断研究与应用。电子鼻由气敏传感器、信号处理和模式识别系统等功能器件组成。电子舌是用类脂膜作为味觉传感器，以类似人的味觉感受方式检测味觉物质。着重阐述了电子鼻与电子舌技术的结构组成，重点介绍了其在食品新鲜度检测、果蔬成熟度评价及饮料、酒类识别等轻工业中的应用现状与发展趋势，并指出了这些信息新技术实现过程中所需要解决的问题。     

免耕播种机播种质量红外监测系统设计与试验

为了对播种机播种质量进行监测,实现精量播种,该文采用红外检测技术,设计了一种对射式传感器,研制了配套播种质量监测系统,并进行了田间播种作业试验研究。该监测系统由多路传感器、监控终端、播种传动传感器组成,结构合理,安装方便,对播种过程出现的堵塞、漏播、缺种等现象,可发出声光报警信号,及时指示故障的位置。田间试验中监测数据与实际测量对比结果表明:该监测系统能够适应田间工作环境,播种质量监测准确率可达95%以上,能够解决人工难以实时监测播种质量的问题,可避免漏播断条现象发生,提高播种机的工作效率和播种质量。     

鱼粉品质检测电子鼻传感器阵列的多特征数据融合优化

为了提高基于仿生嗅觉的鱼粉品质检测装置的鉴别能力,该文利用自主设计的仿生嗅觉鱼粉品质检测装置,提取鱼粉样本的响应特征信息,对其传感器阵列进行多特征数据融合优化。依据各传感器对样本的响应曲线,提取传感器特征值(10×6个)构成原始特征矩阵,后对传感器阵列特征值进行归一化处理,以紧凑性作为评价特征选择方法合理性的标准,采用了3种单特征排序方法(MIC、χ^2、F-test),3种多特征排序方法(RF、LR、SVM),4种特征递减消除方法(RFRFE、SVMRFE、DTRFE、LRRFE)对不同品质的鱼粉进行分类准确率检验,得到基于随机森林的特征递减消除算法(RFRFE)的紧凑性最好,此时最佳的分类准确率为98.3%,特征数目为33个。优化后的传感器阵列特征发生了明显的变化,传感器阵列由原来的10个变为了8个,去掉了传感器TGS2620和传感器TGS2600,特征值也减少了45%。为了避免选择偏差,采用了10折交叉验证方法,再次得到了RFRFE算法具有更佳的紧凑性。该特征选择方法为利用仿生嗅觉技术鉴别其他动物源性原料样本的特征优化提供了新的方法和参考。     

基于绿色农业的智能喷药机器人的设计与应用

随着农业生产的发展，传统的喷药方式存在着药剂浪费、环境污染和作物质量下降等诸多问题。通过智能喷药机器人的应用，提高农业喷药的效率、减少药剂的使用量和环境污染，以推动绿色农业发展和提高农产品质量。结合物联网、大数据和人工智能等技术，设计了一款由底盘、四轮驱动机构、自动配药装置、喷药装置及智能控制器、传感器等组成的喷药智能机器人，能在不同路况下行走工作。通过图像识别技术深度分析农作物所受病虫害情况，通过控制模块获取速度传感器、压力传感器及流量传感器瞬时数值动态调整药液浓度、喷药压力，实现喷药过程平稳工作。该系统的设计为后续根据作物生长态势精准靶向喷药及远距离喷药等功能实现提供思路。     

基于全方位视觉传感器的农业机械定位系统

利用全方位视觉传感器能提供传感器周围360°范围内信息的图像,通过全方位图像可准确获取物体方向角,基于该特点研发一种可以实现农业机械自动导航的定位系统。该系统硬件主要由全方位视觉传感器、人工标识、PC、PC软件和作业车构成。该计算算法首先通过全方位视觉系统的校正,得出相机的成像投影中心;通过图像处理提取人工标识的位置,估算出标识相对于相机的方向角,然后根据圆周角定理和几何关系计算出相机的位置来完成农业机械的定位。在面积为5×3 m2的水泥地面和白炽灯光的实验室中进行试验,平均选取10个位置进行试验,试验结果显示估算点和选取点间距离的平均误差小于3 cm。研究证明该定位系统是可行的。     

吹扫式仿生嗅觉检测装置的设计与性能试验

为研究仿生嗅觉检测装置的检测性能,该文研制了一种吹扫式仿生嗅觉检测装置,选取传感器阵列灵敏度、响应时间和恢复时间作为装置性能指标,分析了传感器加热电压、待检气体湿度以及流量等控制参数对嗅觉检测性能的影响,并进一步验证优化控制参数的仿生嗅觉检测装置对气味检测的可行性和检测性能。试验结果表明:传感器阵列灵敏度随加热电压增大而增大,其响应和恢复时间随加热电压增大而减小,加热电压为5.0 V时,装置性能较佳,其传感器阵列灵敏度分布范围为2.260~4.823,响应和恢复时间分布范围分别为46~53 s、44~70 s;增加气体湿度会使传感器阵列灵敏度减小,同时也会延长响应和恢复时间,气体相对湿度为30%时,装置性能较佳;传感器阵列灵敏度随载气流量的增加先递增后减小,其响应和恢复时间随流量的增加先减小后递增,流量为100 m L/min时,装置性能较佳,其传感器阵列灵敏度分布范围为2.853~7.559,响应和恢复时间分布范围分别为35~50 s、30~50 s;在优化控制参数下,待检气体体积分数范围控制在0.002%~0.020%时,装置检测灵敏度较高,范围为3 577.1~6 700.7;线性特性和重复性较好,决定系数范围为0.901~0.997,变异系数范围为0.832%~9.696%,能满足仿生嗅觉气味检测的要求,可为后续开展仿生嗅觉的应用性研究提供数据参考与技术支撑。     

嗅觉可视化技术及其对四种食醋的识别

不同的食醋散发的香味具有一定的相似性,同时又有各自的特征,怎样快速表达这些特征是一大难点.本研究根据Kenneth S.Suslick教授的思路,筛选出23种卟啉类化合物和7种疏水性pH指示剂,制成一个5x6气体可视化传感器阵列.文章详细介绍了嗅觉可视化系统及其操作步骤,并利用气体可视化传感器阵列对4种传统食醋进行了测试.聚类分析表明,当取相似度为12时,样本被区分为镇江香醋、山西老陈醋、四川麸醋、浙江玫瑰米醋.研究表明嗅觉可视化技术可用于食品气味的可视化分析和图像指纹表达.     

基于近红外光谱技术的淡水鱼品种快速鉴别

为探索淡水鱼品种的快速鉴别方法,该文应用近红外光谱分析技术,结合化学计量学方法,对7种淡水鱼品种的判别分类进行了研究。采集了青、草、鲢、鳙、鲤、鲫、鲂等7种淡水鱼,共665个鱼肉样品的近红外光谱数据,经过多元散射校正(multiplicative scatter correction,MSC)、正交信号校正(orthogonal signal correction,OSC)、数据标准化(standardization,S)等20种方法预处理,在1 000~1 799 nm范围内分别采用偏最小二乘法(partial least square,PLS)、主成分分析(principal component analysis,PCA)和BP人工神经网络技术(back propagation artificial neural network,BP-ANN)、偏最小二乘法和BP人工神经网络技术对7种淡水鱼原始光谱数据进行了鉴别分析。结果表明,近红外光谱数据,结合主成分分析和BP人工神经网络技术建立的淡水鱼品种鉴别模型最优,模型的鉴别准确率达96.4%,对未知样本的鉴别准确率达95.5%。模型具有较好的鉴别能力,采用该方法能较为准确、快速地鉴别出淡水鱼的品种。     

蛋鸡生产参数自动监测系统设计与实现

为了在蛋鸡自然状态下,实现对其采食量、饮水量、排泄量、产蛋时间、蛋质量等生产参数的长期自动采集,克服人工采集工作量大,鸡应激反应造成误差大等问题,设计并实现了基于传感器网络的蛋鸡生产参数自动监测系统。该装置运用人机工程学的设计方法,将传统鸡笼与传感器有机结合,实现了对蛋鸡生产参数无干扰远程监测。该文重点分析了鸡笼的机械结构设计、传感器通信原理、数据采集与分析利用问题。试验结果表明,系统运行稳定,采食量、饮水量、排泄量、蛋质量平均相对误差均小于0.2%。该研究提高了监测数据的精确性,为蛋鸡养殖中科研数据收集、分析和利用提供了新方法和手段,为进一步研究蛋鸡生长过程中生产参数的无干扰自动采集技术提供了参考。     

A neighborhood median weighted fuzzy c-means method for soil pore identification

Complex soil pore geometry and heterogeneity determines the ability of a soil to retain moisture and conduction. These soil properties are widely recognized as key factors of essential ecological functions and services. However, until recently, the existing pore identification methods have the problems of low identification accuracy and operating efficiency, which has restricted the development of soil science. The objective of this study was to propose a neighborhood median weighted fuzzy c-means method based on grayscale-gradient feature (NMFCM-G) to identify soil pore structure automatically and accurately. By combining three-dimensional (3D) printing technology with X-ray computed tomography technology, a 3D simulation model with known aperture (10-cm inner diameter) was adopted to evaluate the pore identification error rate of the NMFCM-G method quantitatively. Compared to the methods commonly applied in previous studies, the NMFCM-G method had the smallest average area relative error (2.98%), which was only 1/6 of that of the Image J method with the largest area relative error (18.46%). The NMFCM-G method also had the smallest average perimeter (5.46%), about 3/5 of that of the Image-Pro Plus method with the largest perimeter relative error (8.35%). Meanwhile, the NMFCM-G method was successfully tested on undisturbed cylindrical soil samples, providing encouraging results in terms of identifying irregular pore structure from the complex hierarchical organization of soil. The results show that the NMFCM-G method had the smallest distribution entropy (0.81), the smallest inter-class correlation (0.164 0), and the largest distribution coefficient (0.11), which proves that the NMFCM-G method performed the best in identifying different soil pore structures. Overall, the NMFCM-G method provides new insights into identifying pore structures and thus could provide an automatic and high-efficiency technique for studying the effects of tillage and freeze-thaw cycles on pore structure and soil quality in the future.     

Colorimetric sensor array for soft drink analysis

Fourteen commercial soft drinks have been analyzed using colorimetric sensor arrays made from a set of 25 chemically responsive dyes printed on a hydrophobic membrane. Digital imaging of the dye array before and after immersion provides a color change profile as a unique fingerprint for each specific analyte. The digital data library generated was analyzed with statistical and chemometric methods, including principal component analysis (PCA) and hierarchical clustering analysis (HCA). Facile identification of all of the soft drinks was readily achieved using comparison of the color change profiles or a PCA score plot. Using a HCA dendrogram, the misclassification rate was <2%, and even very similar sodas were easily differentiated. In addition, the monitoring of soft drinks as they degas or upon dilution also proved to be possible. This work demonstrates the potential of our colorimetric sensor array technology for quality assurance/control applications of sodas and perhaps other beverages as well.     

Determinants of adolescents' soft drink consumption

OBJECTIVE: To identify determinants of adolescents' consumption of carbonated soft drinks (regular and diet), both of total consumption and of consumption at school. DESIGN/SETTING/SUBJECTS: Regular and diet soft drink consumption was measured by food frequency questions that were dichotomised. Several potential environmental and personal determinants of consumption were measured. A total of 2870 (participation rate: 85%) 9th and 10th graders, within 33 Norwegian schools, participated in the study. Multilevel logistic regression analyses were preformed for total soft drink consumption (twice a week or more vs. less) and for consumption at school (once a week or more vs. less). RESULTS: A total of 63% and 27% of the participants reported to drink respectively regular and diet soft drinks twice a week or more, and 24% and 8%, respectively, reported to drink soft drinks once a week or more at school. Preferences, accessibility, modelling and attitudes were the strongest determinants of both regular and diet soft drink consumption. In addition, gender, educational plans and dieting were related to both total soft drink consumption and consumption at school. Pupils with longer distance from school to shop and those in schools with rules concerning soft drink consumption tended to have lower odds of drinking both regular and diet soft drinks at school. CONCLUSION: This study shows that gender, educational plans, dieting, accessibility, modelling, attitudes and preferences all seem to be strong determinants of adolescents' soft drink consumption. Parents and the home environment appear as great potential intervention targets.     

免疫算法在土壤资源质量评价中的应用

Based on the geographic information system （GIS） technology, ArcInfo software was adopted to collect, process and analyze spatial data of Guangdong Province for an evaluation of soil resource quality. The overlay analysis method was used in combining evaluation factors of Guangdong soil resource quality to determine the evaluation units. Because of its favorable convergent speed and its ability to search solutions, the immune algorithm was applied to the soil resource quality evaluation model. At the same time, the evaluation results of this newly proposed method were compared to two other methods： sum of index and fuzzy synthetic. The results indicated that the immune algorithm reflected the actual condition of soil resource quality more exactly.     

冬小麦抗倒春寒性能鉴定方法和指标的比较

倒春寒是限制黄淮冬麦区小麦稳产和丰产的主要因素之一,选育抗倒春寒小麦新品种是降低倒春寒危害的重要手段,建立简单易行、易推广且全面系统的小麦抗倒春寒性能鉴定方法和指标体系对保障粮食安全有着重要意义。本文系统总结了冬小麦抗倒春寒鉴定的方法和指标,进行了比较和评价,并提出鉴定原则和程序。（1）鉴定方法方面,主要包括直接鉴定法、间接鉴定法和综合鉴定法等,直接鉴定法有田间自然鉴定法（引种栽培应用最广,但周期长、工作量大）、人工模拟气候室法（周期短、可重复性强,但对设备与技术要求较高）和生长恢复法（冬前冻害鉴定较多,抗倒春寒鉴定方面较少,一般与其它鉴定方法配合使用）等;间接鉴定法有助倒春寒危害温度阈值的研究,但目前指标与抗寒性关系还未有明确定论,应用受到限制;综合鉴定法鉴定结果全面而准确,但对技术、成本要求较高且工作量较大。因此,在鉴定过程中要结合实际情况合理搭配、综合运用。（2）鉴定指标方面,主要包括形态与农艺、生理生化、气象与生态指标等,形态与农艺指标一般用于田间自然鉴定法,常以数量、质量、颜色和形态变化等作为参考项;生理生化指标一般用于间接鉴定法,包括活性氧、渗透调节物质、内源激素含量、抗氧化酶活性以及光合与呼吸特性等;气象与生态指标常应用于各种鉴定方法,气温、地温、冠层温度、半致死温度和恢复生长最低温度常作为鉴定指标,而光照、CO2、湿度和肥力等因素会影响结果准确性。（3）关于鉴定原则与程序,本文提出了鉴定结果的地域性、指标选用的针对性、鉴定方法与指标的综合性和实用性等鉴定原则,并以易灾期、品种敏感期、敏感器官、鉴定方法和指标、低温程度、田间/盆栽试验、数学分析方法、抗寒性能评价与分级为主要流程的鉴定程序。最后,从抗寒生理研究、新兴技术鉴定研究与应用和鉴定产品的研发等方面做了展望。     

Fabrication and optimization of a conducting polymer sensor array using stored grain model volatiles

During storage, grain can experience significant degradation in quality due to a variety of physical, chemical, and biological interactions. Most commonly, these losses are associated with insects or fungi. Continuous monitoring and an ability to differentiate between sources of spoilage are critical for rapid and effective intervention to minimize deterioration or losses. Therefore, there is a keen interest in developing a straightforward, cost-effective, and efficient method for monitoring of stored grain. Sensor arrays are currently used for classifying liquors, perfumes, and the quality of food products by mimicking the mammalian olfactory system. The use of this technology for monitoring of stored grain and identification of the source of spoilage is a new application, which has the potential for broad impact. The main focus of the work described herein is on the fabrication and optimization of a carbon black (CB) polymer sensor array to monitor stored grain model volatiles associated with insect secretions (benzene derivatives) and fungi (aliphatic hydrocarbon derivatives). Various methods of statistical analysis (RSD, PCA, LDA, t test) were used to select polymers for the array that were optimum for distinguishing between important compound classes (quinones, alcohols) and to minimize the sensitivity for other parameters such as humidity. The performance of the developed sensor array was satisfactory to demonstrate identification and separation of stored grain model volatiles at ambient conditions.     

Spectrophotometric determination of ascorbic acid in canned fruit juices, cordials, and soft drinks with iron(III) and 1,10-phenanthroline as reagents

A simple and accurate spectrophotometric method has been developed for the determination of ascorbic acid in canned fruit juices, cordials, and soft drinks, based on the reduction of iron(III) by ascorbic acid to iron(II), which is then complexed with 1,10-phenanthroline. Background correction is necessary for most samples and can be achieved by copper(II)-catalyzed oxidation of the acid. The calibration graph was linear from 0 to 8 micrograms/mL of ascorbic acid with a slope of 0.12/ppm. The precision for the determination of ascorbic acid in a lemon drink containing 210 micrograms/mL of the acid was 0.9%. Many ingredients commonly found in fruit juices, cordials, and soft drinks do not interfere; however, tannic acid, pyrogallol, and sulfite interfere with the method. A wide range of samples was analyzed for ascorbic acid content by the proposed method. The samples included mango and lemon tea drinks and also grapefruit juices, for which no background correction is needed.     

3YDX-3型烟草打顶抑芽机设计

为改变中国烟草手工打顶、打顶抑芽分步操作的落后局面,开发了3YDX-3型烟草打顶抑芽机。它主要有吹风系统、烟草高度识别系统、打顶系统、抑芽剂喷施系统、刀片消毒装置等组成,能够同时完成烟草打顶、抑芽剂喷施、刀片消毒3项作业。整体机架由液压油缸控制升降,滑动支架与整体机架滑动连接,使机组满足不同生长高度、不同种植行距的烟草使用。烟草高度识别系统、抑芽剂喷施系统采用机电一体化技术,实现了烟草高度的精确识别、准确打顶和对靶施药。试验结果表明,生产率为0.48 hm2/h时,该烟草打顶抑芽机工作安全稳定,打顶高度仿形,抑芽剂喷施准确,能够满足使用要求。     

节水灌溉工程模糊神经网络综合评价模型研究

以模糊理论与神经网络相结合，提出了节水农业灌溉工程模糊神经网络综合评价方法，该方法网络识别率超过传统的方法，克服了模糊系统中的精度问题，其效果又比仅使用人工神经网络系统理论系统的优越。结合河南某节水农业试范区，进行了节水灌溉工程模糊神经网络综合评价分析，评判结果合理、方法简便实用。