视频抄表无线通讯系统

【作者】 孙云龙；

【导师】 常存；

【作者基本信息】 黑龙江大学 ， 物理电子学， 2007， 硕士

【摘要】 仪器仪表被广泛地应用于各行各业的测量系统中,由于某些仪器仪表并没有数据传送的接口,对其进行现场数据读取既浪费人力资源又难以避免人为因素的干扰,因此本系统要实现仪表数字的远程读取。系统的实现需要利用视频技术、图像处理技术、图像识别技术以及无线传输技术。基于现代抄表的现状以及发展趋势,采用数字识别的方法,使用ARM系统搭建软硬件平台,把采集的图像数据按照GPRS网络协议要求打包,采用MA8-9i模块发送到处理中心,处理中心计算机采用BP神经网络实现数字识别。采用带有USB接口的摄像头进行图像采集,省去采集卡,降低了成本,采用ARMSYS2410开发板把采集装置的体积减小化。对仪表的数字图像,运用Visual C++开发工具实现了图像锐化、图像去噪、图像分割、图像特征提取等图像预处理的相关算法。在图像二值化方面,采用了给定阈值的二值化方法,并给出了处理结果。图像的锐化采用了梯度锐化的方法,可以使图像模糊的边缘变得清晰。图像噪声的去除采用的是去除离散杂点的方法,有效的去除了噪声。采用改进边缘检测算法来分割图像,取得了良好的效果。采用13点特征提取法提取数字特征,并加以保存以便送给BP网络进行训练,BP神经网络对于Arial字体的数字可以达到90%以上的识别,提高了识别率。系统减少了对仪表读数的人为因素干扰,也达到远程抄表的目的。本系统集采集、无线传输和图像识别于一体,实现了对仪表数字的远程读取功能,具有实用价值。最后,对本文的工作进行了总结,提出了存在的问题和进一步改进的方向。更多还原

【Abstract】 Instruments are widely applied in the way of the measurement system for many fields. It is a waste of the human resources to read the data of the instruments. The interference of the human factors cannot be avoided during the measurement. The reason is that some instruments have no interface to send the data. So, the purpose of this system is to realize the remote reading. The system will be realized with the video technique, the image processing technique, the image recognition technique and the wireless transmission technique.The dissertation is based on the current situation and the exploitation trend of the modern meter reading and the digital recognition method. The ARM system was used to build the software and the hardware platform and to pack the collected data according to the GPRS network protocol. The MA8-9i module can be used to transmit the data to the processing center. The BP neural network is used to realize the digital recognition. In this dissertation, the USB camera with the image is adopted to save the acquisition card and cut down the cost, the ARMSYS2410 exploitation board is used to reduce the volume of the acquisition device. In order to deal with the digital image of the measuring appliance, the Visual C++ exploitation device is adopted to realize the relevant algorithm of the image pretreatment such as the image sharpening, the image denoising, the image segmentation, and the extraction of the image characteristics. The processing results were given by using of the threshold method in respect of the image binaryzation. The method of gradient sharpening method is used to make the fuzzy edge change to clear. The method of removing the discrete mixed spot is used to remove the image noise. The method that improves the edge examination algorithm is utilized to segment the image and the good effect has obtained. The 13 points characteristics extraction method is employed to obtain and save the digital characteristics in order to send them to the BP network to take the training. The recognition of the BP neural network for the Arial typeface may reach up to 90%, which has improved the recognition rate. This system can reduce the interference of the human factors for the meter reading, and it also can achieve the purpose of remote meter reading. The system is the integration of the sampling, the wireless transmission and the image recognition. It realizes the function of the remote meter reading and has the practical value. In the last content in the dissertation, the work of dissertation was summarized, and the ameliorative direction was also proposed.更多还原