节点文献

基于图像的水位自动检测研究

【作者】 刘治锋

【导师】 任明武;

【作者基本信息】 南京理工大学 , 模式识别与智能系统, 2004, 硕士

【摘要】 当前检测水位的方法有许多种,如浮子式水位计、压力式水位计、气介或液介超声波水位计等,在不同的情况下有着不同的应用。远程水情监控系统中由于环境等因素影响,常用的水位检测方法已经不能适用,为解决该问题提出了基于图像的水位自动检测技术,该技术是远程水情监控系统的核心技术之一,是非接触式水位自动检测。该技术基于图像处理的方法,丰富了图像处理内容同时为水位自动检测提供了新的思路。 本文解决的关键技术问题和提出的方法如下: (1)面向具体环境,规划、设计了水位检测的整体流程。 (2)提出了基于图像水位自动检测的方法。为达到自动检测目的,对采集的水位图像进行了一系列处理,包括形状校正、光强校正、目标分析、滤除无用信息、水位区域定位和水位线精确检测等方法。 (3)实现了图像中水位高度和实际海拔水位高度之间的快速转换技术。该技术采用创建查找表方法。 本文提出的水位自动检测技术在山西汾河监控系统中已投入使用。到目前,经过一年多的验证,检测到的水位高度误差在2.0cm以内,其性能不受温度和一定条件下光照的影响,适合于不规则落差、大范围的水位测量,为满足我国偏远地区、特别是山区恶劣条件下的河流湖泊水情监控和水位自动检测,提供了一种新的水位自动检测技术。

【Abstract】 Currently, there are many methods to detect water level, such as float-type water level gauge, pressure-type water level gauge and ultrasonic water level gauge etc. It has been used in different conditions. Auto Water Level Detection (AWLD), based on image, is one of core techniques of Remote Water Monitor System (RWMS). This technique is an untouched-AWLD. Based on image, this technique enrich the knowledge of image processing, and give an novel method of detection water level.The key techniques and new methods presented in this thesis are listed below:(1) For the real and specific AWLD condition, constructed the whole flow chart of recognition.(2) Presented a novel AWLD method based on image. After image rectification and light intensity rectification, the chain code technique was used to trace object’s contour and remove useless information, and the line of water level in the image was located accurately.(3) Presented a solution of transformation between the water level in image and the real altitude water level.The algorithms presented in this thesis have been put into practice in Fenhe RWMS, Shanxi Province. During more than one year, the error rate of detected water level is less than 2.0cm. It hasn’t been affected by temperature and water quality. It can also be used in the situation of abnormal fall and large area water detection. This techniques provide a brand-new auto water level measure for remote zone of our country, especially for stream monitor and auto water level detection in mountainous badly conditions.

  • 【分类号】TP274.4
  • 【被引频次】36
  • 【下载频次】823
节点文献中: