小气候自记图图像数据采集的数字化处理方法

提出一种用数字图像处理技术(digital image processing,DIP)和算法获取小气候自记图图纸上数据的方法,解决了此类图纸中弧形时间轴变成直角坐标轴的问题,并且提高了数据的获取量,基本方法是利用图象软件Imagetool中的“Point”功能获取小气候自记图图象上的记录曲线上点的像素值,然后通过我们给出的算法转换成相应的数据资料,步骤包括：自记图纸扫描、自记图纸的图像纠偏、事件数字化、时间轴的标准化、温湿度的计算、现场时间的计算、时间和数字化图形一致性生成。同时通过实例比较和分析说明了手工与DIP方法在数据取样密度和表现性能差异上的来源。图4参4。

IMAGE-PROCESSING-BASED DIGITIZATION OF MICROCLIMATE EVENT RECORDER CHARTS

A digital transformation and interpretation method of microclimate charts for weekly or daily auto-recorder of relative humidity and temperature is developed. The basic idea is that using the "point" function of software UTHSCSA Imagetool to acquire the pixel values of the event in the scanned chart image and then process these arc-shaped isochronal data with algorithms proposed by us to render these data. The protocols include: chart scanning, image rectification, event digitization with software Imagetool, normalization of time axle, interpretation of humidity, temperature and time in a rectangular coordinates. For a given period of time, the density of sampling with manual method is proportional to the projected length of event curve, while in the digital method it relates immediately to the length of curve. In addition, the resolution of data interval with manual method relies on eyes rather than on the pixel density as in digital image processing method. Therefore, compared to the traditional manual method, the digital image process can greatly improve the data collecting density and efficiency and reveal some delicate and rapid processes of changes. Fig 4, Ref 4