摘要
针对水文测验平台在测流过程中需要人工监视上游水域情况避免水草缠绕在测流悬杆上,工作效率较低的问题,提出了一种基于机器视觉识别水草并向测流系统实时反馈的研究方法。首先总控中心对摄像机视频进行帧化处理,得到水域图片;而后对其进行预处理,在过滤颜色特征及部分图像噪声的同时保留水草的形态特征,降低后期数据处理量;接着采用改进的阈值处理方法,自适应选取局部阈值减弱亮度不均匀的影响,实现水草等前景区域的分割;在二值图基础上进行形态学操作,强化前景区域的形态特征;最后通过面积过滤,排除水藻等杂物干扰,实现水草的识别。实验结果表明,该方法可以有效识别复杂水体背景下的水草,识别流程平均耗时0.64s,能够满足悬杆测流系统自适应调节对信号的实时要求,极大地降低了人工劳动强度,为实现全自动化无人测流夯实基础。
Aiming at the problem that the hydrological test platform needs to manually monitor the upstream waters during the flow measurement process to avoid the aquatic plants winding on the flow measuring suspension rods and its low efficiency,a re-search based on machine vision to identify the aquatic plants and provide real-time feedback to the flow measurement system is pro-posed.First,the master control center frames the camera video to obtain a picture of the water area.Then,it is pre-processed to fil-ter the color features and part of the image noise while retaining the morphological features of the water plants,reducing the amount of post-data processing.Next,an improved threshold processing method is adopted to adaptively select local thresholds to reduce the influence of uneven brightness,and realize the segmentation of foreground areas such as aquatic plants.Perform mor-phological operations on the basis of binary images to strengthen the morphological characteristics of the foreground area.Final-ly,filter by area to eliminate the interference of algae and other sundries,and realize the identification of aquatic plants.The ex-perimental results show that this method can effectively identify aquatic plants under complex water background,and the average time of identification process is 0.64s,which can meet the real-time requirements of adaptive adjustment of suspension rod flow measurement system for signal,greatly reduce the manual labor intensity,and lay a solid foundation for the realization of fully automated unmanned flow measurement.
作者
高霄凡
武利生
张金柱
赵帮强
GAO Xiaofan;WU Lisheng;ZHANG Jinzhu;ZHAO Bangqiang(College of Mechanical and Vehicle Engineering,Taiyuan University of Technology,Shanxi Taiyuan 030024,China)
出处
《机械设计与制造》
北大核心
2025年第11期222-226,共5页
Machinery Design & Manufacture
基金
国家自然科学基金青年项目(51905367)。
关键词
水草
预处理
阈值处理
形态学操作
面积过滤
全自动化无人测流
Aquatic Plants
Pretreatment
Threshold Processing
Morphological Operation
Area Filter
Fully Au-tomated Unmanned Flow Measurement
