基于环境标示物和深度学习的自动定位智能导航系统设计  

Design of an automatic positioning intelligent navigation system based on environmental markers and deep learning

作　　者：贾媛媛 JIA Yuanyuan(Shaanxi Institute of Technology,Xi’an 710300,China)

机构地区：[1]陕西国防工业职业技术学院,西安710300

出　　处：《自动化与仪器仪表》2025年第2期284-288,293,共6页Automation & Instrumentation

摘　　要：针对工业生产环境中自动导引车的自动定位智能导航系统设计,研究在图像分析计算模块中引入了双边分割网络。同时,采用空洞空间金字塔池化层与多尺度特征融合模块对其进行改进,实现图像的语义分割。此外,引入了基于跳点搜索算法的A*算法实现路径规划。结果显示,在自动定位智能导航系统的性能测试中,系统的实测定位点几乎与标定点的运行轨迹完全一致,且导航定位误差均小于6 mm,平均误差值仅为2.25 mm,具有较高精度。说明研究设计的自动定位智能导航系统具有极高的导航定位精度,能够用于实际的导航与路径规划中,对于提升工业生产以及生产环境设计的效率和可靠性具有重要意义。In order to design an automatic positioning intelligent navigation system for guided vehicles in industrial production en-vironments,a bilateral segmentation network was introduced into the image analysis and calculation module.At the same time,the hollow space pyramid pooling layer and multi-scale feature fusion module are used to improve it and achieve semantic segmentation of the image.In addition,A∗algorithm based on skip point search algorithm was introduced to achieve path planning.The results showed that in the performance testing of the automatic positioning intelligent navigation system,the measured positioning points of the system were almost identical to the running trajectory of the calibration points,and the navigation positioning errors were all less than 6 mm,with an average error value of only 2.25 mm,indicating high accuracy.The research and design of an automatic positioning intelligent navigation system has extremely high navigation positioning accuracy and can be used in practical navigation and path plan-ning,which is of great significance for improving the efficiency and reliability of industrial production and production environment de-sign.

关 键 词：环境标识物 深度学习 智能导航 BiSeNet A*算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]