摘要
在强风天气来临时,未关闭的建筑幕墙开启窗极易被吹落,严重危害了人身与财产安全。为了及时准确地发现未关闭的开启窗,提出了一种基于计算机视觉的幕墙开启窗检测方法。广角摄像头从建筑底部仰角拍摄幕墙外立面全景影像,采用语义分割模型U-Net提取外立面掩码图像,并从掩码图像提取用于透视变换的参考点,透视变换法将因仰角拍摄产生的畸变进行校正,校正后的开启窗位置分布规则,可直接按照预设区域准确裁剪出开启窗区域图像,然后采用ResNet-18卷积神经网络分类模型判断开启窗开闭状态。建立开启窗位置与楼宇房间号映射关系表,从而指导责任人员及时定位并关闭开启窗。所提方法在某高层建筑进行了测试。结果表明:所提方法可以在不同光照条件下有效检测到未关闭的开启窗,检测时间优于2 s。可见,所提方法在准确性与及时性上满足实际应用需求。
The opening windows of architectural curtain wall are easily blown off when strong winds come,which seriously endanger personal and property safety.In order to discover the opening windows promptly and accurately,an opening windows detection approach based on computer vision was proposed.The panorama image of curtain wall was captured by wide angle camera which located in bottom of the building,U-Net semantic segmentation model was used to obtain facade mask image,and extract reference points from the mask image which were used in perspective transform,the perspective transform was used to correct the facade distortion which caused by elevation angle shooting,the corrected open windows position distribution were regularly which can be directly cropped by the preset area.Then the ResNet-18 classification model was used to judge the opening and closing status.In order to guide responsible person to locate and close opening windows timely,the mapping relation map between windows positions and room numbers was needed to be established.The proposed method was verified on a high-rise building.The results show that the proposed method has a high level of effectiveness with various brightness images,and the detection time is less than 2 s.Therefore,the proposed method meets application requirements in accuracy and timeliness.
作者
卢佳祁
姚志东
庄浩然
LU Jia-qi;YAO Zhi-dong;ZHUANG Hao-ran(Central Research Institute of Building and Construction(Shenzhen)Co.,Ltd.,Shenzhen 518055,China;Shenzhen Engineering Research Center of Intelligent Inspection for Curtain Wall,Shenzhen 518055,China)
出处
《科学技术与工程》
北大核心
2022年第20期8748-8754,共7页
Science Technology and Engineering
基金
国家重点研发计划(2020YFB2103604)
中冶建筑研究总院重大课题项目(YJA2018Kj01)。
