基于卷积神经网络的嵌入式视觉感知交互系统设计与实现  

Design and Implementation of Embedded Visual Perception Interaction System Based on Convolutional Neural Network

作　　者：陶金 王智勇 林鸿生 周怡伶 TAO Jin;WANG Zhiyong;LIN Hongsheng

机构地区：[1]海军士官学校,安徽蚌埠233012 [2]92682部队,广东湛江524000

出　　处：《科技创新与应用》2024年第3期35-39,共5页Technology Innovation and Application

摘　　要：针对语音智能助理无法提供周围环境的视觉感知问题,该文设计并实现一个视觉感知交互系统。该系统的基本结构由语音识别、语音播放、图像采集、中央处理控制等模块组成,具备语音交互、目标检测等功能。该系统设计选择语音识别专用芯片、利用卷积神经网络技术完成识别,采用基于图分割截块的算法进行目标分割。实验结果表明,系统性能良好,能够实现对周围环境的视觉感知并进行语音交互。In order to solve the problem that the speech intelligent assistant is unable to provide visual perception of the surrounding environment,a visual perception interaction system is designed and implemented in this paper.The basic structure of the system consists of speech recognition,voice playback,image acquisition,central processing control and other modules,with voice interaction,target detection and other functions.In this system,the special chip for speech recognition is selected,the convolutional neural network technology is used to complete the recognition,and the algorithm based on graph segmentation is used to segment the target.The experimental results show that the system has good performance and can realize visual perception of the surrounding environment and voice interaction.

关 键 词：卷积神经网络 视觉感知 嵌入式 语音识别 图分割截块 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]