摘要
针对带有七段码的数字液晶屏,设计了一种基于安卓手机的液晶屏数字识别系统。通过手机拍照获取液晶屏图像,利用安卓本地接口功能调用计算机视觉库,对图像进行处理;针对图像的倾斜问题,通过图像预处理和霍夫变换取得图像倾斜角,并进行倾斜矫正;利用直方图对数字字符进行分割,以网格法提取字符的特征值,用三层BP神经网络进行液晶屏数字字符的识别。实际检测结果表明识别率很高,而且识别速度很快,该系统基于安卓手机,使用方便,便于携带,可以实时操作。
In view of liquid crystal display with 7-segment, a LCD digital recognition system based on Android phone is designed. The picture is gained from the Android phone and processed with open source computer vision library which can be used with the method of Android Java native interface. The inclination angle of the picture is fixed and adjusted through the preprocess and Hough Transform in terms of the inclined picture. The numerical character is extracted with histogram and its characteristic value is collected with grid method. Three-layer BP neural network is selected to recognize the numerical characters. The result shows that the recognition rate is high. The system is based on Android phone so that it′s very convenience and easy to carry and can realize real-time operation.
出处
《微型机与应用》
2015年第17期36-38,42,共4页
Microcomputer & Its Applications
关键词
安卓
数字识别
开源计算机视觉库
BP神经网络
Android
digital recognition
open source computer vision library
BP neural network
