图像识别技术在水位监测中的比测分析

Comparative Analysis of Water Level Monitoring Technology Using Image Recognition

摘要: 受河床下切影响，现有水位监测的自记水位井和压力式水位计存在着枯期记录不到数据的情况，而采用传统人工水尺读数耗时且效率不高，不符合水文信息化、智慧化发展的方向。为解决该问题，引入图像识别水位系统并对其进行了比对分析。本次比对按照《水位观测标准》(GB/T50138-2010)要求，收集了人工比测数据和浮子式水位计比测数据，计算了相关不确定度。经过一系列比对分析，证明该系统能够满足水位监测的要求，同时符合水文信息化发展方向。

Abstract: At present, the water level monitoring is affected by the incision of the river-bed. The existing self-recording water level wells and pressure-type water level meters have no data recorded during the dry period. Using the traditional artificial water gauge to read the water level is time-consuming and inefficient, which is unbeneficial to the development of hydrological information and intelligent development. In order to solve this problem, an image recognition water level system was introduced, which can automatically read and upload water level data in real time. The manual data and float type water level meter data are collected in accordance with the requirements of “Water Level Observation Standard (GB/T50138-2010)” and the uncertainty of these data is analyzed. Through comparative analysis, it is proved that this system can be in accord with the requirements of water level monitoring and the development of hydrological informatization.