基于K-means算法的二步迭代道路检测算法  

Two-step Iterative Road Detection Algorithm Based on K-means Algorithm

作　　者：史迪玮 毛剑琳[1] SHI Di-wei;MAO Jian-lin(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《软件导刊》2018年第9期110-114,共5页Software Guide

摘　　要：为增强现有基于机器视觉的自动导引车AGV对非结构化路面的适应能力,提出一种基于K-means聚类分析算法的二步迭代道路检测算法。算法实现了自动选择路面样本区域与自主动态添加路面样本,克服了基于其它机器学习算法的道路识别方法需要人工收集大量路面样本进行训练的缺陷。实验仿真结果显示,该方法能有效降低光照、阴影、车道线等对道路识别的影响,能够适应含有多种不同障碍物的道路场合。This paper proposes a two-step iterative road detection algorithm based on K-means clustering analysis algorithmto enhance the adaptability of current automated guided vehicles （AGV） based on machine vision to unstructured pavement.Thisalgorithm implements the automatic selection of pavement sample areas and autonomous dynamic addition of pavement samples.However other road recognition methods based onmachine learning algorithmsneed to manually collect a large number of pavement samples for training,this algorithm has overcome the defects.The simulation results show that this method can effectively reduce the influence of illumination,shadow and lane line on road recognition,and it can adapt to roadsituation with many different obstacles.

关 键 词：非结构化路面 路面检测 聚类分析 二步迭代 机器视觉 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]