摘要
在基于摄像机的场景测量中,图像采集的同时需要获得摄像机的参数,即进行摄像机定标。在利用参照物的自动标定过程中,首要任务就是实现测量场景中参照物(或模板)的自动检测与精确定位。为此提出了一种复杂背景下模板检测与定位的新方法。该方法包括模板检测与精确定位两个步骤。前者利用图像分割与形态学方法实现模板的粗定位;后者则结合边缘检测和Radon变换等技术确定模板的精确位置。实验结果表明,该方法实现简单、性能可靠,可适用于各种复杂的场景图像中。
In the process of camera-based scene measurement the camera parameters should be obtained at the same time of image acquisition, i. e. performing the camera calibration. During the automatic calibration process based on the reference object the first task is to detect automatically and locate precisely the reference (or template) in realizing detection scene. For this purpose a novel method for template detection and location under complex background is provided in this paper, which consists of two steps, i.e. template detection and accurate location. The former implements the template coarse location using the images segmentation and morphological algorithm and the later fixes the template accurate position combined its edge detection and radon transform techniqnes. The experimental result shows simple implement and reliable performance of proposed method, which can be applied in any complex scene images.;
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2004年第1期87-90,共4页
Systems Engineering and Electronics
基金
国家自然科学基金资助课题(60202004)
关键词
模板检测
精确定位
摄像机
图像测量
template detection
accurate location
camera
image measurement
