U.S.-Turkey Cooperative Research: Food Quality and Safety by Kernel Classification

项目描述：该项目支持明尼苏达州明尼阿波利斯大学电气工程系Ahmed Tewfik博士与土耳其安卡拉比尔肯特大学电气与电子工程系Enis Cetin博士领导的团队之间的合作研究。他们建议通过冲击声学分析来研究污染或缺陷核的检测和分类方法的发展。pi计划开发基于冲击声学分析的污染或缺陷核检测和分类，或核撞击坚硬板时产生的声音，因为这种声音激励模式很容易适应高通量分类系统。该方法在榛子、小麦和玉米上的应用证明了其有效性，可以实现对大量食品的非侵入式快速检查，从而在食品安全方面取得突破。尽管食品警报的质量控制越来越严格，但目前的方法大多是基于对某些选定食品的侵入性化学分析或根据食品的颜色进行分类。虽然化学分析能给出最准确的结果，但不可能分析大量的食品或将其应用于某些食品。随着计算机技术的进步，信号处理技术变得越来越有吸引力，现在可以将低成本、非侵入性的信号处理系统集成到食品供应链中，从而提供质量控制。该方法使用了语音和说话人识别技术的大量扩展，以处理在撞击瞬间不同核方向所导致的可变性。该项目更具体地侧重于检测榛子中的黄曲霉毒素（一种致癌物质）和虫洞、被昆虫破坏的麦粒以及玉米中的黄曲霉毒素和虫洞。用高效液相色谱法对检测和分类结果进行验证。范围：该项目将导致开发实时的、经济上可行的系统，以去除受黄曲霉毒素污染的完整玉米和榛子仁或受昆虫感染的小麦仁；为土耳其和美国的本科生和研究生以及食品行业提供信号处理、农业工程和分子生物学方面的跨学科培训项目；学生参与数据收集和分析；还有算法开发。经过一些修改，本项目开发的算法和方法也可以用于其他坚果和农产品，如杏仁和无花果。所开发的算法和其他基础知识可能适用于一般的多元信号处理，而不是声学。土耳其在榛子生产和加工方面拥有丰富的专业知识，而美国是谷物生产和加工专业知识的来源。研究团队反映了该项目的跨学科性质，一些研究人员为该项目带来了强大的语音分析和处理背景，并将这些方法交叉到食品安全领域。

0352965 TewfikDescription: This project supports collaborative research between Dr. Ahmed Tewfik, Department of Electrical Engineering, the University of Minnesota, Minneapolis, Minnesota and a team headed by Dr. Enis Cetin Department of Electrical and Electronics Engineering, Bilkent University, Ankara, Turkey. They propose to study the development of methods for detection and classification of contaminated or defective kernels by using analysis of impact acoustics. The PIs plan to develop contaminated or defective kernel detection and classification based on the analysis of impact acoustics, or the sound created when a kernel strikes a hard plate, as this mode of acoustic excitation is easily adapted to high throughput sorting systems. Demonstrating the effectiveness of the method in applications to hazelnuts, wheat, and corn, could achieve a breakthrough in food safety by enabling the non-invasive, rapid inspection of large quantities of food items. Despite increasingly rigorous quality control stand in food alerts, current approaches are mostly based on invasive chemical analysis of some selected food items or sorting food items according to their color. Although chemical analysis gives the most accurate results, it is impossible to analyze large quantities of food items or to apply it to certain items. Signal processing techniques have become attractive with the advances in computer technology, and it is now possible to integrate low cost, non-invasive signal processing systems providing quality control into the food supply chain. The approach uses substantial extensions of speech and speaker recognition techniques to deal with the variability that results from different kernel orientation at the instant of impact. The project focuses more specifically on the detection of aflatoxins (a carcinogenic material) and insect tunnels in hazelnuts, wheat kernels damaged by insects, and aflatoxins and insect tunnels in corn. Detection and classification results will be verified with high performance liquid chromatography. Scope: The project will lead to the development of real-time, economically feasible, systems to remove whole intact corn and hazelnut kernels that are contaminated with aflatoxins or wheat kernels infested with insects; interdisciplinary training programs for undergraduate and graduate students and the food industry in Turkey and the USA in signal processing, agricultural engineering and molecular biology; the involvement of students in data gathering and analysis; and algorithm development. With some modifications, the algorithms and methods developed in this project can be also used for other nuts and produce such as almonds and figs. The algorithms and other fundamental knowledge to be developed may be applicable to multivariate signal processing in general, beyond acoustics. Turkey has a wealth of expertise in hazelnut production and processing, while the United States is the source of expertise in grain production and processing. The research team reflects the interdisciplinary nature of the project, and several of the investigators bring to the project a strong background in the analysis and processing of speech and will crossover such methods to food safety.