仪表显示的图像识别算法研究.docVIP

PAGE PAGE 25 仪表显示的图像识别算法研究 摘要：随着社会的逐渐发展，人类的生活越来越趋于智能化。本文根据当今社会对于图像识别研究的发展现状，针对目前人们生活中人工读表的弊端，提出了通过采集仪表显示的图像并进行图像识别算法处理来达到智能自动读表的方法。 为了能快速获得采集数据，减少人们生活中繁复的人工作业。本文通过多样的图像处理来代替人眼识别图像。只需要得到采集到的图像，就可以利用计算机来进行计算和识别，得出最后的数字。本文采用了一系列的图像处理方法，包括图像的去噪，二值化分割，边缘检测和基于数学形态学的膨胀腐蚀操作等。同时通过多种尝试和比较各种方法的优缺点得到了一套简易而又完善，快速的图像识别算法。 在进行多次测试试验后，本文采用数码相机来进行图像的采集，同时经过图像预处理、图像分割、图像识别等一系列流程得出了较为完善的图像采集和识别系统，为未来信息传递智能化提供了基础，对于促进工业发展或是改善生活水平都有重要的意义。 关键词：图像预处理、二值化、边缘检测、形态学、去噪、图像分割、图像匹配 The research of image recognition displayed by the instrument Abstract: With the continuous development of society, peoples lives become more and more intelligent. Based on the current development in todays society for the study of image recognition, according to the present disadvantages of manual meter reading in people’s lives, this page proposed the way by collecting the instruments display image and then deals it with image recognizing algorithms to achieve intelligent automatic meter. In order to quickly gather data, reducing manual work in people’s lives complicated. The page uses a series of image processing to replace human eye image recognition. Just need the collected images, we can use a computer to calculate and identify, then we will arrive at a final figure. We used a variety of image processing methods, including image denoising, thresholding segmentation, expansion of edge-detection based on mathematical morphology and corrosion and so on. And though many try and compare the disadvantages and advantages of each methods, we get a simple, improved and fast image recognition algorithms. After a series of testing, we decide to use digital cameras for image capture, and then combined with image processing, image segmentation, image recognition, and other processes to conclude a more perfect image acquisition and recognition system. This system can provide the basis for future information delivery intelligence, for both the promotion of industrial development and improved living standards has important