摘要
为了进一步满足高速公路养护及管理技术需求,本文针对高速公路路面破损特征复杂性、图像噪声严重等难点问题,首先以差影法理论为基础,提出了一种自适应确定路面图像阈值的方法,为后续路面破损特征检测与识别提供条件;同时,应用改进的中值滤波算法、多阈值平均的图像分割方法及BP神经网络完成了破损路面特征的去燥、边缘平滑、图像分割以及精准识别,有效提升了破损路面检测的效率和质量。
in order to further meet the technical requirement of highway maintenance and management,this paper aiming at the complexity and image noise of the damagecharacteristicsof highway pavement,first of all,base on the subtraction method theory to proposed a adaptivemethod which determine the threshold of pavement image,to provide conditions for the detection and recognition of the damages; at the same time,and applied the improved median filtering algorithm,multi-threshold average image segmentation and BP neural network to complete the dry,smooth edge and image segmentation of the damaged pavement characteristics,then effectively enhance the detection efficiency and quality of damaged pavement.
出处
《自动化与仪器仪表》
2016年第10期131-132,共2页
Automation & Instrumentation
关键词
高速公路
路面破损
中值滤波
图像分割
highway
pavement distress
median filter
image segmentation
