摘要
为了解决传统水果秤称重、支付方式的不足,在Jetson Nano平台驱动CSI摄像头进行实时图像检测,通过基于TensorFlow搭建的卷积神经网络模型完成水果识别,从而实现水果的自动称重计价播报。通过语音交互的方式了解用户的购买需求后,根据计价结果显示二维码进行扫码支付,整个购买过程无须售货员介入,并且实现了无接触购物。实验结果证明,所设计的图像识别水果秤可在2 s内完成水果的识别、称重和计价。水果识别准确率达到90%以上,同时实现了不同水果混装情况的辨别。
In order to solve the shortcomings of traditional fruit scale weighing and payment methods,the Jetson Nano platform drives the CSI camera to perform real-time image detection,and completes fruit recognition through the convolutional neural network model built on TensorFlow,thereby realizing the automatic weighing and pricing of fruits. After understanding the user’s purchase needs through voice interaction,the QR code is displayed according to the pricing result to scan the QR code. The entire purchase process does not require the intervention of the salesperson,and contactless shopping is realized. Experimental results prove that the designed image recognition fruit scale can complete the process of fruit recognition,weighing and pricing within 2 seconds. The fruit recognition accuracy rate reaches more than 90%,and at the same time,it realizes the discrimination of different fruit mixing situations.
作者
许龙铭
麦启明
卢家俊
陈苇浩
XU Longming;MAI Qiming;LU Jiajun;CHEN Weihao(Communication Engineering College,Guangzhou City Institute of Technology,Guangzhou 510800,China)
出处
《电子设计工程》
2022年第6期174-178,共5页
Electronic Design Engineering
