| 赵宇飞,刘彪,王毅,孟亮,刘必旺.基于数字图像处理的土石坝坝料合格性智能检测方法[J].水利学报,2022,53(10):1194-1206 |
| 基于数字图像处理的土石坝坝料合格性智能检测方法 |
| Intelligent detection method for material qualification of earth-rock dam based on digital image processing |
| 投稿时间：2022-04-15 |
| DOI：10.13243/j.cnki.slxb.20220285 |
| 中文关键词: 土石坝料级配检测 数字图像处理 SIFCM算法 级配修正模型 坝料合格性检测 |
| 英文关键词: material gradation detection of earth-rock dam digital image processing SIFCM algorithm gradation correction model dam material qualification detection |
| 基金项目:中国水科院三型人才基金项目（GE0145B042022） |
|
| 摘要点击次数: 1540 |
| 全文下载次数: 171 |
| 中文摘要: |
| 土石坝坝料的合格性检测通常是通过判断现场筛分试验获得的级配参数是否满足设计要求来实现的，然而通过筛分试验获取级配参数的方法存在采样率低，操作过程繁琐，智能感知程度差等缺点以致检测结果代表性差。为了提高坝料级配参数的智能检测程度，依托辽宁某电站现场挖坑检测位置处的图像和级配数据，采用融合空间信息的直觉模糊C均值聚类（SIFCM）算法进行土石坝料数字图像的分割，并利用等效椭球体积的方法实现了土石坝料的三维体积重构，进一步通过基于BP神经网络的级配修正模型修正后，得到真实条件下的坝料全级配特征曲线，进而获得评价坝料合格性的4个指标：最大粒径、P5含量、曲率系数Cc和不均匀系数Cu。实际工程应用表明，本文建立的基于SIFCM_BP算法的坝料级配特征智能识别修正模型具有较高的识别精度，本文方法为大坝碾压施工前坝料合格性快速判别与施工过程中坝料压实特性的实时评价提供了重要支撑。 |
| 英文摘要: |
| The qualification testing of earth-rock dam materials is usually realized by judging whether the gradation characteristic parameters obtained from on-site screening test meet the design requirements. However, the method of obtaining the gradation characteristic parameters through the test has some shortcomings, such as low sampling rate, cumbersome operation process and poor intelligent perception, resulting in poor representativeness of the testing results. In order to improve the intelligent detection of dam material gradation parameters, relying on the images and gradation data at the test location of one pumped storage power station in Liaoning Province, the intuitionistic fuzzy C-means clustering algorithm fused with spatial information (SIFCM) is used to segment the image of earth-rock dam materials. Next, the 3D volume reconstruction of earth-rock dam material is achieved by the equivalent ellipsoidal volume method. Then the gradation characteristic curve of dam material under real conditions is obtained through the gradation correction model based on BP neural network. Finally, four evaluation indexes of dam material qualification are obtained:maximum particle size, P5 content, curvature coefficient Cc, and uneven coefficient Cu. The practical engineering application shows that the intelligent identification and correction model of dam material gradation characteristics based on the SIFCM_BP algorithm established in this paper has high identification accuracy. The method in this paper provides an important support for the rapid identification of dam material qualification before the compaction construction and the real-time evaluation of dam material compaction characteristics during construction. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|