基于OpenCV的手持式和面板式数字仪表自动采集系统

摘要

For the purpose of accomplishing the character recognition, the following steps are to be taken: First, an image/video sensing device is used to acquire the image in the reading area of a portable and digital panel meter based on OpenCV Computer Vision. After geometry correction, adaptive thresholding and pre-processing of the image, Hough Transform is applied to correct the sloping font before splitting up the image vertically then horizontally to realize the image normalization. At last, train the 12 eigenvalues of the seven-segment code to constitute the neural network to ultimately realize the character recognition.%利用OpenCV计算机视觉库，通过图像/视频感知设备对手持式和面板式数字仪表的显示区域进行图像采集，在对图像几何校正、自适应阈值二值化等预处理后，使用霍夫变换校正倾斜字体，随后对图像先垂直再水平分割，进行图像归一化处理，利用七段码的12个特征值经过训练构成神经网络，最终完成对字符的识别。