基于边缘检测法的风自记纸图像数字化  

Digitization of Wind Autographic Records of EL Type Based on Edge-detection Algorithm

作　　者：李亚丽 黄少平[3] 鞠晓慧[4] LI Ya-li;HUANG Shao-ping;JU Xiao-hui(Shaanxi Provincial Meteorological Information Center,Xi’an 710014,China;Key Laboratory of Eco-environment and Meteorology for Qinling Mountains and Loess Plateau of Shaanxi Provincial Meteorological Bureau,Xi’an 710016,China;Jiangxi Provincial Meteorological Information Center,Nanchang 330046,China;National Meteorological Information Center,Beijing 100081,China)

机构地区：[1]陕西省气象信息中心,陕西西安710014 [2]陕西省气象局秦岭和黄土高原生态环境气象重点实验室,陕西西安710016 [3]江西省气象信息中心,江西南昌330046 [4]国家气象信息中心,北京100081

出　　处：《计算机技术与发展》2022年第1期160-164,169,共6页Computer Technology and Development

基　　金：中国气象局山洪地质灾害防治气象保障工程项目(2017-04);陕西省自然科学基础研究计划项目(2020JQ-977);陕西省气象局秦岭和黄土高原生态环境气象重点实验室开放研究课题基金(2019Z-6)。

摘　　要：风自记纸是客观记录风向风速的珍贵观测资料。利用数字图像处理技术实现风自记纸迹线数据的提取和图像的数字化,是对气象历史档案的有效拯救,也是获取长序列历史分钟、小时等高分辨率风数据的有效途径,这些数据必将为极端天气事件成因和气候长期变化趋势原因分析提供有利条件。应用基于Otsu算法改进的Canny图像边缘检测方法设计开发了EL型电接风自记纸迹线数据提取软件系统,将读取的彩色风自记纸图像进行灰度处理,采用Canny算法对图像进行边缘检测,通过Radon变换求解图像的旋转角度对图像进行倾斜校正,采用Otsu算法对灰度图进行二值化,然后根据风自记纸的特征,从二值图中提取风速风向迹线,并经过去除噪音、细化、去毛刺等一系列处理获取风向风速网格和迹线坐标,根据风速风向计算方法将迹线数据转换生成分钟、小时和日最大风向风速。应用效果表明:系统对标准的EL型电接风自记纸图像识别率可达到100%,经与A文件中人工整理的风向风速记录对比,能够满足风自记纸图像数字化的数据质量和精度需求。Wind autographic recording paper recorded objectively wind speed and direction is valuable meteorological observation data.Using digital image processing technology to digitize wind autographic records of EL type is an effective test for rescuing historic meteorological archives and is also an effective way to get high resolution wind data such as long sequence of historical minute and hourly data.These data will certainly provide favorable conditions for the analysis of the causes of extreme weather events and long-term climate change trends.By the improved Canny image edge-detection method based on Otsu algorithm,a digital processing system of wind autographic recording paper is designed and developed.Firstly,the color images are grayed and carried on edge detection by Canny operator,and the angle of image rotation can be obtained by Radon Transform to correct the inclined image.Secondly,we can extract the wind speed curve and wind direction vertical bar from the binary image enhanced by means of Otsu algorithm to get optimal thresholds for image segmentation and get the final coordinate points list of wind direction,wind speed grid and track coordinates after a series of processing,such as noise removal,thining and deburring.Finally,we get the real data of one minute,hours and days maximum wind speed and wind direction according to the projection analysis and wind speed and direction calculation method.The system application results prove that the digital processing system of wind autographic recording paper meets basically the quality and precision requirements of data extraction and the image recognition rate of normative wind autographic recording paper image can reach 100%.By comparing with the wind direction and speed records calculated manually in file A,the average deviation and root mean square error are all within acceptable range.

关 键 词：风自记纸图像 Canny图像边缘检测 最大类间方差法 迹线坐标数据提取 图像数字化 

分 类 号：TP39[自动化与计算机技术—计算机应用技术] P413.21[自动化与计算机技术—计算机科学与技术]