手持式马铃薯品质多通道离散光谱检测装置设计与试验

Design and Experiment of a Handheld Multi-Channel Discrete Spectrum Detection Device for Potato Processing Quality

马铃薯是我国第四大主要粮食作物,随着马铃薯主食化战略的提出,其市场占比逐年攀升,但各地甚至同区域内马铃薯品质参差不齐,严重影响了马铃薯行业的发展。实现马铃薯品质快速无损检测对马铃薯主食化产业的发展有着重要的现实意义。该研究以研发低成本马铃薯品质无损快速检测装置为目的,采用连续投影算法(SPA)分析光谱仪环境下马铃薯加工品质特征波长的分布情况,根据标准正态变换(SNV)预处理状态下的模型结果选取了一个包含7个波段(700,750,800,850,900,950和1000 nm)的多通道光谱传感器,并根据马铃薯特殊的表皮特征及内部质地均匀性,设计了一种手持式马铃薯多品质可见/近红外局部漫透射检测装置。利用研发装置建立了马铃薯多品质偏最小二乘预测模型,马铃薯干物质含量、淀粉含量预测模型验证集均方根误差分别为1.05%和1.02%。同时,基于QT的开发工具,采用C语言编写了实时分析设备控制软件,实现了对马铃薯内部品质的一键式实时无损检测。对研发装置检测稳定性和精度进行了试验验证。结果表明,研发的手持式马铃薯多品质传感器检测装置可以满足现场实时检测需求,为马铃薯主食化产业的发展提供技术支撑。

Potatoes are China’s fourth major food crop.With the government’s continuous emphasis on potato production and sales and the strategy of potato staple foods,its market share has also increased year by year.However,the quality of potatoes varies from place to place and even in the same region,which has a serious impact.The development of the potato industry.Therefore,realizing rapid non-destructive testing of potato quality has important practical significance for developing the potato staple food industry.This paper aims to develop a low-cost non-destructive rapid detection device for the potato quality.The successive projections algorithm(SPA)is used to analyze the distribution of characteristic wavelengths of potato processing quality in a spectrometer environment.According to the model results in the preprocessing state of the standard normal variate transformation(SNV),a selection of 7 bands(700,750,800,850,900,950 and 1000 nm)is selected.Based on the potato’s special epidermal characteristics and internal texture uniformity,a handheld potato multi-quality visible/near infrared local diffuse transmission detection device is designed,including a light source module,a spectrum acquisition module,a control module,and a power supply.Module,display module,the size of the whole device is 140 mm×80 mm×70 mm,and the weight is 340 g.A potato multi-quality partial least squares prediction model was established using the research and development equipment.The root means square error of the potato dry matter and starch content prediction model verification sets were 1.05%and 1.02%,respectively.At the same time,based on the development tools of QT,the real-time analysis equipment control software was written in C language,which realized the one-click real-time non-destructive inspection of the internal quality of potatoes.Finally,the testing stability and accuracy of the R&D device were tested and verified.The results show that the developed handheld potato multi-quality sensor detection device can meet the needs of on-site real-time detection and provide technical support for developing the potato staple food industry.