摘要
提出了一种基于检测和验证相结合的指针综合提取方案,目的是实现特种车辆指针式仪表图像识别的自动识别和自动矫正。该方案优化利用了Hough变换法、极坐标中心投影以及边缘点集合的最小二乘法等算法,使得指针的自动提取无论针对何种品质的图像都非常准确。经验证,采用该方法进行特种车辆仪表指针提取,准确度达到99%以上。
A detection and verification based approach is applied to pointer exact extraction for automatical calibration and automatic recognition of dial instruments for vehicle car.The synthetic approach optimizes algorithms such as Hough Transformation,polar central projection,Least Square algorithm,etc.,which brings the high performance to dial reading results.It is tested that,the veracity of the synthetic approach can reach above99%.
作者
郭迎春
张伏龙
张丹阳
白真龙
王涛
GUO Yingchun;ZHANG Fulong;ZHANG Danyang;BAI Zhenlong;WANG Tao(Special Vehicle Institute, Beijing 100072;Academy of Armored Forces Engineering of PLA,Beijing 100072;School of Automation,University of Science and Technology, Beijing 100083)
出处
《太赫兹科学与电子信息学报》
2017年第2期273-278,共6页
Journal of Terahertz Science and Electronic Information Technology
关键词
指针式仪表
指针提取
检测与验证
dial instrument with pointer
pointer extraction
detection and verification
