基于Hough-Harris的消音壁顶点检测  

Soundless Wall Vertex Detection Based on Hough-Harris

作　　者：李国旺 李英[1] 马韵琪 夏晨旭 LI Guowang;LI Ying;MA Yunqi;XIA Chenxu(School of Opto-Electronic Engineering,Changchun University of Science and Technonlgy,Changchun 130022)

机构地区：[1]长春理工大学光电工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2023年第2期106-113,共8页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：吉林省科技厅重点研发项目(20200401070GX)。

摘　　要：在消音壁领域中,消音壁的顶点检测是后续对消音壁毁伤评估、三维重建等任务的关键,检测出的顶点准确率直接影响后续任务的精确度。为提高检测出消音壁顶点的准确率,提出一种基于Hough-Harris的消音壁顶点检测算法。首先,使用Hough算法检测出的角点作为基点;然后,利用Harris检测出消音壁边框上的角点,选取以基点步长为10个像素的领域内角点作为样本点;最后,建立回归模型,利用高斯提出的最小二乘法求出最小误差,预测出消音壁顶点坐标。实验结果表明,本文算法预测顶点坐标与真实顶点坐标偏差量都小于Hough算法和Harris算法检测出的角点,IoU评估指标高达98%以上,高于Hough算法和Harris算法的IoU指标。In the field of deadening wall,the detection of deadening wall vertices is the key to subsequent tasks such as damage assessment and 3D reconstruction of deadening wall.The accuracy of detected vertices directly affects the accuracy of subsequent tasks.In order to improve the accuracy of detecting deadening wall vertices,this paper proposes a deadening wall vertices detection algorithm based on Hough-Harris.Firstly,t he corner point detected by Hough algorithm is used as the base point.Then,Harris was used to detect the corner points on the bezel of the silenced wall,and the corner points in the field with a base point step of 10 pixels were selected as sample points.Finally,the regression model was established,and the least square method proposed by Gauss was used to calculate the minimum error and predict the vertex coordinates of the silenced wall.Experimental results show that the deviation between the predicted vertex coordinates of the proposed algorithm and the real vertex coordinates is smaller than the corner points detected by Hough algorithm and Harris algorithm.The IoU evaluation index of the proposed algorithm is higher than 98%,which is higher than the IoU index of Hough algorithm and Harris algorithm.

关 键 词：HOUGH算法 HARRIS算法 顶点检测 消音壁 

分 类 号：TP274.3[自动化与计算机技术—检测技术与自动化装置]