摘要
基于人工智能的卷积神经网络框架,对电表采用轮廓分析进行特征提取,实现0到9的电表数字识别。步骤是先通过设备拍摄电表读数,获得图像后对图像进行灰度化和二值化处理,然后进行字符分割、图像识别和相应的特征值提取,最后识别出电表的读数。
Based on the convolutional neural network framework of artificial intelligence,the contour analysis is used to extract the features of the meter,and digital recognition from 0 to 9 of the meter is realized.The step is to take the reading of the meter through the equipment,obtain the image,then grayscale processing and binary processing are carried out for the image,and carry out the character segmentation,image recognition and the corresponding feature value extraction,and finally recognize the reading of the meter.
作者
丁宸炀
马晓普
叶菁
Ding Chenyang;Ma Xiaopu;Ye Jing(School of software,Nanyang Normal University,Nanyang,Henan 473061,China)
出处
《计算机时代》
2020年第4期94-96,100,共4页
Computer Era
基金
南阳师院大学生实践创新项目(SPCP2019231)。
关键词
人工智能
抄表
卷积网络
分割
artificial intelligence
meter reading
convolutional network
segmentation
