湖南电力 ›› 2025, Vol. 45 ›› Issue (3): 35-41.doi: 10.3969/j.issn.1008-0198.2025.03.005

• 源网协调与能源转换利用 • 上一篇    下一篇

多设备融合的水工建筑物水下检测技术

张亦可1,2, 张军1, 莫剑3, 张学武4, 张振5   

  1. 1.国网湖南省电力有限公司电力科学研究院,湖南 长沙 410208;
    2.长沙理工大学水利与环境工程学院,湖南 长沙 410114;
    3.国网湖南省电力有限公司,湖南 长沙 410004;
    4.河海大学,江苏 南京 210000;
    5.国网湖南省电力有限公司水电分公司,湖南 长沙 410004
  • 收稿日期:2025-01-06 发布日期:2025-07-02
  • 作者简介:张亦可(1992),男,工程师,主要从事水电站设备设施电状态评价、故障诊断工作。
  • 基金资助:
    国家自然科学基金项目(61671202); 国网湖南省电力有限公司科技项目(5216A5220020)

Underwater Detection Technology of Hydraulic Structure With Multi-Equipment Integration

ZHANG Yike1,2, ZHANG Jun1, MO Jian3, ZHANG Xuewu4, ZHANG Zhen5   

  1. 1. State Grid Hunan Electric Power Company Limited Research institute, Changsha 410208, China;
    2. School of Hydraulic and Environmental Engineering, Changsha University of Science and Technology,Changsha 410114, China;
    3. State Grid Hunan Electric Power Company Limited, Changsha 410004, China;
    4. Hohai University, Nanjing 210000, China;
    5. State Grid Hunan Electric Power Company Limited Hydropower Branch, Changsha 410004, China
  • Received:2025-01-06 Published:2025-07-02

摘要: 针对水下大坝复杂环境中缺陷检测精度低的问题,提出一种基于多传感器数据融合的缺陷检测方法,结合声呐设备和光学摄像头,利用卡尔曼滤波对声呐数据进行去噪,并采用限制对比度的自适应直方图均衡算法(CLAHE)增强光学图像对比度,通过加权最小二乘法实现两种数据的融合,生成包含全局轮廓与高清细节的综合图像。基于卷积神经网络(CNN)对融合图像进行处理,实现裂缝和沉积物等缺陷的精准识别和分类。实验结果显示,该方法在被试大坝中成功检测出3处裂缝及2处沉积物堆积区域,可显著提升缺陷检测的精度和可靠性。

关键词: 水下结构, 多传感器融合, 缺陷检测, 数据去噪, 卷积神经网络

Abstract: To solve the problem of low accuracy of defect detection in the complex underwater dams environment, a defect detection method based on multi-sensor data fusion is proposed. By combining the sonar equipment and optical camera, the sonar data is denoised by Kalman filter, and the optical image contrast is enhanced by the contrast limited adaptive histogram equalization(CLAHE) algorithm. At the same time, the weighted least square method is used to achieve the fusion of the two kinds of data, and the comprehensive image containing the global contours and high-definition details is generated. Then, convolutional neural network(CNN) is used to process the fusion image to realize accurate identification and classification of defects such as cracks and sediments. The experimental results show that the method successfully detected 3 cracks and 2 sediment accumulation areas in the test dam, significantly improving the accuracy and reliability of defect detection.

Key words: underwater structure, multi-sensor fusion, defect detection, data denoising, convolutional neural network

中图分类号: