Department of Electronic, Electric and Automatic Control Engineering, Universitat Rovira i Virgili, 43007 Tarragona, Spain.
Sensors (Basel). 2019 Apr 11;19(7):1741. doi: 10.3390/s19071741.
This paper presents a color-based classification system for grading the ripeness of fruit using a battery-less Near Field Communication (NFC) tag. The tag consists of a color sensor connected to a low-power microcontroller that is connected to an NFC chip. The tag is powered by the energy harvested from the magnetic field generated by a commercial smartphone used as a reader. The raw RGB color data measured by the colorimeter is converted to HSV (hue, saturation, value) color space. The hue angle and saturation are used as features for classification. Different classification algorithms are compared for classifying the ripeness of different fruits in order to show the robustness of the system. The low cost of NFC chips means that tags with sensing capability can be manufactured economically. In addition, nowadays, most commercial smartphones have NFC capability and thus a specific reader is not necessary. The measurement of different samples obtained on different days is used to train the classification algorithms. The results of training the classifiers have been saved to the cloud. A mobile application has been developed for the prediction based on a table-based method, where the boundary decision is downloaded from a cloud service for each product. High accuracy, between 80 and 93%, is obtained depending on the kind of fruit and the algorithm used.
本文提出了一种基于颜色的分类系统，用于使用无电池近场通信 (NFC) 标签对水果的成熟度进行分级。该标签由一个连接到低功耗微控制器的颜色传感器组成，微控制器连接到 NFC 芯片。标签由商业智能手机产生的磁场能量供电，智能手机用作读取器。颜色计测量的原始 RGB 颜色数据转换为 HSV（色调、饱和度、值）颜色空间。色调角和饱和度用作分类的特征。为了展示系统的稳健性，比较了不同的分类算法来对不同水果的成熟度进行分类。NFC 芯片的低成本意味着具有感应能力的标签可以经济地制造。此外，如今，大多数商业智能手机都具有 NFC 功能，因此不需要特定的读取器。在不同的日子里对不同的样本进行测量，用于训练分类算法。训练分类器的结果已保存到云端。已经开发了一个移动应用程序，用于基于基于表的方法进行预测，其中边界决策是为每个产品从云服务下载的。根据水果的种类和使用的算法，获得了 80%到 93%的高精度。
