加油站中基于自适应阈值分割的车辆停车检测方法  

VEHICLE PARKING DETECTION METHOD BASED ON ADAPTIVE THRESHOLD SEGMENTATION IN GAS STATION

作　　者：肖军弼[1] 朱风珍 钱志军 程千才 Xiao Junbi;Zhu Fengzhen;Qian Zhijun;Cheng Qiancai(College of Computer and Communication Engineering, China University of Petroleum(East China), Qingdao 266580, Shandong, China;China Petroleum Planning and Engineering Institute, Beijing 100089, China)

机构地区：[1]中国石油大学(华东)计算机与通信工程学院,山东青岛266580 [2]中国石油天然气有限公司规划总院,北京100089

出　　处：《计算机应用与软件》2019年第2期252-257,共6页Computer Applications and Software

基　　金：赛尔网络下一代互联网技术创新项目(NGII20160106);中国石油天然气有限公司规划总院基于物联网技术的加油站智能总线项目

摘　　要：针对加油站复杂环境下难以获取停车数据的问题,提出一种基于自适应阈值分割的车辆停车实时检测方法。以加油位为检测对象,使用改进的均值背景法更新加油位的背景;将背景差分法与Otsu算法相结合,提出一种改进的自适应阈值分割算法;设计停车检测算法,检测出加油车辆的停车时间点、车牌号以及停车持续时间。实验结果表明,该算法能有效检测出车辆的停车数据,并能过滤掉过往车辆和行人等噪声干扰。Aiming at the problem that it is difficult to obtain parking data in complex environment of gas station, we proposed a real-time detection method of vehicle parking based on adaptive threshold segmentation. Taking the refueling position as the detection object, we used the improved mean background method to update the background of the refueling position. Combining background difference method with Otsu algorithm, an improved adaptive threshold segmentation algorithm was proposed. We designed a parking detection algorithm to detect the parking time, license plate number and parking duration of refueling vehicles. The experimental results show that the algorithm can effectively detect the parking data of vehicles and filter out the noise interference of passing vehicles and pedestrians.

关 键 词：加油站 停车检测 均值背景法 OTSU算法 自适应阈值分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]