摘要
基于字符形态特征提出一种快速的机械仪表判读算法.该方法中采用平滑滤波对灰度化后的图像去除噪音,再使用OTSU算法进行二值化,同时根据图像投影确定表盘区域,最后将Canny算子和Hough变换结合对图像进行矫正并分割字符.字符判读模块中对图像进一步去除噪音,对字符形态特征进行加权计算和对比,准确识别出仪表示值.通过实验证明,该算法辨识率高,有较好的实用性.
We presents a fast interpretation algorithm based on morphological features of characters.Smoothing filters to remove noises on grayscale images,and binarized grayscale images used by OTSU algorithm.After the binary operation,the dial region was found through image projection.We extracted edges using the Canny operation,corrected images with Hough transform,and splitted numeric characters from images.In the character interpretation module,this algorithm opened and closed operations to remove noises deeply.I...
出处
《中国计量学院学报》
2010年第4期337-341,共5页
Journal of China Jiliang University
基金
浙江省科技厅重大科技专项(No.2007C13091)
关键词
机械式仪表
形态特征
自动判读
图像处理
mechanical digital meter
morphological feature
auto-recognition
image processing
