基于ARM的嵌入式图像识别检测系统的设计

Design of embedded image recognition system based on ARM

摘要:

图像识别的主要目的是使用计算机作为工具对目标图像进行处理、解析与应用,通过数据分析检测出具有不同特征的目标和对象,发展至今其已成为了人工智能的基础。本文基于ARM嵌入式芯片提出了一种结合尺度不变特征变换匹配算法的图像识别检测系统,系统硬件部分采用模块化设计的思想以提高系统的兼容性,分为图像获取、数据采集、数据存储、图像识别等模块；软件部分采用斑点检测匹配匹配算法进行图像识别以提高图像识别速度与精度。实验结果表明所设计的系统具有识别速度快、精度高、可靠性高、故障少的特点。

Abstract:

As one of the most important fields of artificial intelligence, image recognition is a technology that uses the computer to process, analyze and understand the image. In this paper, an image recognition and detection system based on ARM embedded chip is proposed, The hardware of the system adopts the idea of modular design to improve the compatibility of the system, it is divided into image acquisition, data acquisition, data processing and image recognition module; the software part of the scale invariant feature transform matching algorithm in order to improve the image recognition and image recognition speed and accuracy. The experimental results show that the designed system has the characteristics of high speed, high precision, high reliability and low failure.