【摘要】 为了实现摄像头采集万用表表头图像的识别,研究了一种针对万用表数字图像的自动识别方法。首先对图像进行预处理,包括提取数字区域、提取区域图片灰度化、平滑处理。然后采用局部自适应二值化方法对图形进行二值化,以及对二值化图像进行形态学处理,并采用投影法对字符串进行分割。最后应用穿线法对数字进行识别,有效降低了误识率。实验结果表明,该方法对数字万用表识别效果好,快速简单。更多还原

【Abstract】 In order to recognize the images that the camera collected, this paper study a method which can realize the digital image’s automatic recognition. First of all, preprocessing the images that it’s collected. It includes extraction of digital regional, image graying and smoothing. Then using local adaptive binarization method to deal with graphics, as well as to the binary image morphology processing, and using the projection method to split the string. Finally using threading method to identify digital, effectively reduce the rate of course. The experimental results show that the method of digital multimeter recognition effect is good,fast and simple.更多还原