基于计算机视觉的汽车仪表指针检测方法  

Novel automobile meter pointer detection algorithm based on computer vision

作　　者：张冀[1,2] 王俊宏[1] 尉迟明[1] 丁明跃[1] 侯文广[1] 

机构地区：[1]华中科技大学生命科学与技术学院,湖北武汉430074 [2]天津医科大学生物医学工程学院,天津300070

出　　处：《计算机工程与科学》2013年第3期134-139,共6页Computer Engineering & Science

基　　金：国家973计划资助项目(2011CB933103);中英国际科技合作项目(2009DFA12290)

摘　　要：最常用的汽车仪表为指针式仪表,工业上汽车仪表指针的检测主要是对指针角度和指针导光性能的检测。汽车仪表检测的准确性是汽车质量控制的关键。目前我国汽车行业普遍采用人工仪表检测方式,劳动强度大,质量控制不稳定。根据指针式汽车仪表图像的特点,采用最小灰度相减法从两张背景相同而仪表指针位置不同的图像中提取到仪表指针,然后运用二值化、细化和Hough变换识别出指针角度。同时,利用提取的指针实现对指针的导光性能检测。实验结果表明,算法具有鲁棒性强、准确性高、实时性好等优点,已成功应用于基于计算机视觉的汽车仪表自动检测系统。The most common automobile meter is the pointer meter. The pointer detection of the au- tomobile meter in industry is mainly concentrated on the detection of the pointer angle and the measure ment of the light guide property of pointer. Accurate detection is the key of automobile quality. Current- ly, manual meter detection is widely used in Chinese automobile industry, which is tedious and with a high variability of quality control. According to the image characteristics of automobile pointer meter, a minimum gradient comparing subtraction method is used to extract the meter pointer from two images with the same background and pointers in different positions. Then, the pointer angle can be recognized by the binary method, the thinning algorithm and the Hough transform. Meanwhile, the extracted pointer is utilized to realize the detection of pointer light guide property. Experimental results show this proposed method is robust, accurate and can be performed in real-time. It has been successfully applied to an automatic detection system of automobile meter based on computer vision.

关 键 词：汽车仪表 指针检测 导光 最小灰度相减 HOUGH变换 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]