改进ORB-SLAM在嵌入式系统中实现  

Improved Implementation of ORB-SLAM in Embedded System

作　　者：王磊[1] 杨永夫 潘明然 郭宏林 WANG Lei;YANG Yongfu;PAN Mingran;GUO Honglin(College of Mechanical Engineering,Shenyang Ligong University,Shenyang 110159,China;Liaoshen Industrial Group Co.,Ltd.,Shenyang 110045,China)

机构地区：[1]沈阳理工大学机械工程学院,沈阳110159 [2]辽沈工业集团有限公司,沈阳110045

出　　处：《组合机床与自动化加工技术》2024年第10期37-41,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：辽宁省教育厅基本科研项目(LJKFZ20220186);沈阳市中青年科技创新人才项目(RC200537)。

摘　　要：针对飞行载体的实时ORB-SLAM实现问题,提出一种在嵌入式系统实现的改进ORB(oriented FAST and rotated BRIEF)单目视觉里程计算法。算法首先对输入图像进行灰度化、高斯滤波预处理实现简化运算和图像去噪,考虑到算法移植及在嵌入式系统实现,将图像预处理和ORB图像特征提取与匹配等功能封装为IP(intellectual property)核,布置到硬件系统中,提高特征提取与匹配的速度和正确率,保证位姿估计实时性。搭建ZYNQ嵌入式系统,开展对比实验,实验结果表明:改进后的算法特征点匹配率提高了3.78倍,特征提取与匹配的耗时缩短为原来的1/8,处理图像的帧率可以达到19 fps,满足实时性要求。In response to the real-time ORB-SLAM implementation challenges on aerial platforms,an improved Embedded System implementation of the ORB(oriented FAST and rotated BRIEF)monocular visual odometry algorithm is proposed.The algorithm first preprocesses raw images through grayscale conversion and Gaussian filtering to simplify computations and remove noise.With a focus on algorithm portability and embedded system implementation,functions such as image preprocessing,ORB feature extraction and matching are encapsulated into an intellectual property(IP)core and deployed on hardware systems to enhance feature extraction and matching speed and accuracy,ensuring real-time pose estimation.An embedded system based on ZYNQ is constructed,and comparative experiments are conducted.The experimental results demonstrate that the improved algorithm achieves a 3.78-fold increase in feature point matching rate,reduces the time consumption for feature extraction and matching to 1/8 of the original,and achieves a frame rate of 19 fps for image processing,meeting real-time requirements.

关 键 词：ORB-SLAM 特征提取与匹配 ZYNQ 嵌入式 FPGA 

分 类 号：TH165[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]