摘要
本文提出一种基于卷积神经网络模型的嵌入式手写数字识别的方案。该方案以卷积神经网络为手写数字识别模型结构,使用MNIST手写体训练集训练网络模型,以STM32H750XBH6为嵌入式终端设备芯片,搭载RT-Thread物联网操作系统,利用RT-Thread AI kit将手写数字识别模型集成到板级支持包(Board Support Package,BSP),实现嵌入式手写数字的识别。
A scheme of embedded handwritten digit recognition based on convolutional neural network model is proposed. The scheme uses convolutional neural network as the structure of handwritten digit recognition model, uses MNIST handwriting training set to train the network model, uses STM32H750XBH6 as the embedded terminal device chip with RT-Thread IoT operating system, and integrates the handwritten digit recognition model into BSP through RT-Thread AI kit to realize embedded handwritten digit recognition.
作者
王冲
耿玉菊
刘光伟
张博文
张贝宁
WANG Chong;GENG Yuju;LIU Guangwei;ZHANG Bowen;ZHANG Beining(Department of Mathematics and Computer Science,Hengshui University,Hengshui Hebei 053000,China)
出处
《信息与电脑》
2022年第8期54-57,共4页
Information & Computer
基金
2021年河北省大学生创新创业训练项目“基于STM32的数字识别研究与应用”(项目编号:s202110101017)。
关键词
卷积神经网络
手写数字识别
嵌入式人工智能
convolutional neural network
handwritten digit recognition
embedded artificial intelligence