摘要
针对目前直升机桨叶图像中圆形标记点检测和识别难的问题,本文结合桨叶圆形标记点排布规律,提出了一种快速检测算法。首先,采用CLAHE和中值滤波对图像进行预处理,增大图像对比度和信噪比;然后,根据圆的几何性质对封闭边缘进行初步筛选,获取图像中所有类圆轮廓;最后,依据桨叶上粘贴的圆形标记点呈阵列结构排布规律,滤除背景中的干扰,获得桨叶标记点圆心。通过对多幅桨叶图像中圆形标记点进行检测和识别,验证了该方法具有自适应能力强、速度快、精度高的优点,已用于直升机高速旋转桨叶运动轨迹和形变的高精度视觉测量。
Aiming at the difficulty in detecting and recognizing circular markers in helicopter blade images,this paper proposed a fast detection algorithm based on the regularity of circular marker arrangement in helicopter blades.Firstly,images are preprocessed by the Contrast Limited Adaptive Histogram Equalization(CLAHE)and median filtering to increase the contrast and SNR.Then,according to the geometric properties of circles,closed edges are preliminarily screened to obtain all circular contours in the image.Finally,according to the arrangement regularity of circular marker array structure on the helicopter blade,the interference of the background is filtered out and the centers of all markers are obtained.By detecting and recognizing circular markers in multiple helicopter blade images,the proposed method has the advantages of high adaptability,fast speed,and high precision,.It has been applied in the high precision measurement of the motion trajectory and deformation of helicopter rotor blades rotating at high speed.
作者
张育斌
熊邦书
陈垚锋
程起有
ZHANG Yubin;XIONG Bangshu;CHEN Yaofeng;CHEN Qiyou(Science and Technology on Rotorcraft Aeromechanics Laboratory,China Helicopter Design and Research Institute,Jiangxi 333000,China;Key Laboratory of Jiangxi Province for Image Processing and Pattern Recognition,Nanchang Hangkong University,Nanchang 330063,China)
出处
《中国体视学与图像分析》
2021年第4期428-434,共7页
Chinese Journal of Stereology and Image Analysis
基金
航空科学基金(No.20185702002)。
关键词
直升机桨叶
低曝光图像
圆形标记点
阵列结构
helicopter blade image
image of low exposure
circular marker
array structure mode
