|
发布时间: 2019-09-16 |
图像分析和识别 |
|
|
收稿日期: 2018-12-12; 修回日期: 2019-04-17
基金项目: 河南省科技攻关计划项目(172102210050);河南省高等学校重点科研计划项目(15A510003,14B170012);水利部黄河泥沙重点实验室开放课题基金项目(2017001)
第一作者简介:
张来胜, 男, 主要研究方向为数据与可视化研究。E-mail:13592571760@139.com;
秦泽宁, 男, 硕士研究生, 主要研究方向为水文模型分析。E-mail:944536211@qq.com; 刘佳琪, 男, 硕士研究生, 主要研究方向为农业信息化。E-mail:iiwm@vip.qq.com; 陈健, 男, 硕士研究生, 主要研究方向为含沙量测量。E-mail:1306324645@qq.com.
中图法分类号: TP391.4
文献标识码: A
文章编号: 1006-8961(2019)09-1504-10
|
摘要
目的 高填方渠段渗漏监测技术是南水北调工程安全监测的关键技术之一。针对目前高填方渠道渗漏检测易受环境干扰导致判断结果不准确的问题,设计了基于Gabor-SVM(support vector machine)的南水北调中线工程高填方渠道水泥坡面破损识别模型。方法 首先对高填方渠道水泥渠面图像进行预处理;然后研究Gabor小波的多方向/尺度选择特性,对提取的水泥坡面图像特征进行分析,寻找最佳尺度/方向参数组;最后根据训练好的样本特征,用SVM进行水泥坡面破损等级的分类处理,将破损类别划分为正常、裂缝、孔洞及破碎4种。在相同环境下,将Gabor-SVM与直方图-SVM、灰度共生矩阵-SVM、Canny边缘检测算法-SVM的破损识别进行比较分析。结果 基于Gabor-SVM的水泥面破损识别方法在小波取6/12参数时,整体识别结果最佳,其中正常、裂缝、孔洞、破碎的识别率分别为98%、63%、88%、90%,平均识别率约为85%,相比其他几种方法的平均识别率(约50%),Gabor-SVM方法具有更好的识别能力。结论 基于Gabor-SVM的水泥渠面破损识别模型有一定的识别效果,但对裂缝类别的高填方水泥坡面识别效果不理想,还需要进一步深入研究,以期为查找南水北调高填方渠道渗漏隐患提供技术支持。
关键词
南水北调中线工程; 多尺度/方向特征提取; 破损识别; Gabor小波; SVM分类
Abstract
Objective The leakage monitoring technology of high fill channels is key to the safe monitoring of the South-to-North Water Transfer Project. Aimed at addressing the current problem of leakage detection for high fill channels being easily affected by the environment and resulting in inaccurate judgment results, a model based on Gabor-support vector machine (SVM) is designed for damage monitoring in the cement slope of the high fill channel in the middle route of the South-to-North Water Transfer Project. The high fill channel is widely distributed in the middle route of the South-to-North Water Transfer Project. As a result of the high filling height, wide distribution range, and complex engineering geological conditions of the middle route of the South-to-North Water Transfer Project, the lining panel is cracked, the canal slope surface is damaged, and seepage occurs. Although the effect of leakage is small, the long channel still leads to a large leakage. Therefore, monitoring the seepage of the high fill channel is necessary to ensure its safe operation. Method The image of the high fill channel cement is preprocessed. In general, because the acquired image is affected by various noises, it needs to be preprocessed before extracting its features. These processing methods include image enhancement, median filtering, and grayscale processing. Then, Gabor wavelet is used to extract the texture features of the image, as well as the image convolution, commonly used amplitude, and phase to represent the texture features. The amplitude information reflects the energy spectrum of the image and is relatively stable. Therefore, amplitude information is selected as the extracted characteristic data. After analyzing the mean and variance eigenvalues of the amplitude, the image features of the variance are linearly separable. Therefore, the variance characteristic data of amplitude are considered as the characteristic data for classification and recognition. Under the scale and direction of different Gabor wavelets, the image characteristics of the extracted cement slope are analyzed to find the optimal scale and direction parameter group. The scale range is 17, and the direction range is 113. Different Gabor filters are obtained according to different directions and scales. Different Gabor filters are used to filter the high fill cement slope and thereby obtain different image features. Finally, according to the well-trained sample features, SVM is used to classify the damage degrees of cement slope, and the recognition results are refined with the following labels:normal, crack, fracture, and hole. At the same time, various feature extraction methods are studied to objectively reflect the recognition effects of Gabor-SVM. These methods include histogram-SVM, grayscale symbiosis matrix-SVM, and Canny edge detection algorithm-SVM. Result Experimental results show that the damage recognition model of the cement surface based on Gabor-SVM tends to have a stable value when the small wave has the 6th scales and the 12th directions. The recognition rate of the normal slope image is generally good and stable, mostly distributed between 0.8 and 1.0. The recognition rate of the slope image under the crack category presents stable growth from low to high, but the overall recognition rate is low, with most values being in the range of 0.500.65. The recognition rate of the slope image of the hole type fluctuates greatly, and its recognition rate has a significant relationship with scale changes. For example, when the scale value is 1 or 2, the recognition rate is low. When the scale value is 3, 4, or 5, the recognition rate increases gradually. When the scale value is 6 or 7, the recognition rate decreases. The recognition rate is mostly distributed between 0.78 and 0.88. A certain relationship exists between the size of the slope image recognition rate of the fracture category and the scale. It has the characteristic of fluctuating growth from low to high and then to low, and the recognition rate is generally between 0.80 and 0.95. The normal, crack, hole, and fracture recognition rates are 0.98, 0.63, 0.88, and 0.90, respectively. The average recognition rate of the Gabor-SVM method is approximately 0.85. Compared with the average recognition rates of the other methods (approximately 0.50), that of the proposed method has better recognition ability. Conclusion The damage recognition model based on Gabor-SVM has a slight recognition effect. The recognition effect peaks given 6 scales/12 directions. The average recognition rate of the Gabor-SVM method for the slope of the high fill channel cement is approximately 0.85. Meanwhile, the cement surface recognition effect of the crack category is unsatisfactory at 0.63. Thus, further research is required to provide technical support for finding the hidden dangers of high fill channels in the South-to-North Water Transfer Project.
Key words
middle route of South-to-North Water Transfer Project; feature extraction of multiple scales and directions; damage recognition; Gabor wavelet; support vector machine (SVM) classification
0 引言
一般来说,填方高度大于6 m的渠段称为高填方渠道[1]。高填方渠道广泛分布于我国南水北调中线工程[2]。由于南水北调中线工程跨区域广、填方高度大、工程地质条件复杂,高填方渠段会因衬砌面板开裂或渠坡水泥面破损等出现渗(漏)水[3-4]。南水北调中线工程渠线较长,渗漏量较大,一旦长期渗漏就会导致溃堤,直接影响沿线居民的生命财产安全。因此,高填方渠段渗漏监测技术是南水北调工程安全监测的关键技术之一。
目前,由于部分河流渠道、库区堤坝在结构上和高填方渠段有一定相似处,其渗漏检测方法对高填方渠段渗漏检测有一定借鉴意义。国内外对大坝渗漏的检测方法主要有地质雷达探测法、电阻率法、超声波检测法和图像处理检测法[5-7]。渗漏检测的算法模型有渠段渗流模型、多源数据融合模型、多目标逆建模等[8-9]。基于图像处理的方法能检测高填方渠段水泥面破损,可为查找高填方渠道渗漏隐患提供支持,从而有助于高填方渠段的渗漏检测。
图像处理的检测方法已应用于墙面、路面、单板、矿井、隧道等大面积的裂缝识别,处理步骤包括图像滤波、图像分割、特征提取和图像分类等。其中,特征提取是高填方渠段水泥面破损识别的关键环节之一。特征提取方法一般有二值化分割、Canny边缘检测、灰度共生矩阵等方法。灰度共生矩阵分析图像的纹理特征,反映各灰度级在空间上的分布特性,特征分量达14个左右,可得出不同像素相对位置的空间信息[10-12]。虽然一般的图像检测法在大面积检测上具有优势,但难以实现对渗漏区突变的特征进行有效识别。目前需要图像检测法能在实现大面积检测的同时,又能自动识别突变特征的高填方渗漏监测模型[13]。
Gabor小波变换方法对图像边缘特征敏感,能提供良好的方向和尺度选择特性,而且能对光照变化有良好的适应性[14-15]。支持向量机(SVM)分类器根据有限数据样本集得到的判别函数,对独立的测试集仍能够得到较小的误差,具有较强的泛化能力[16-17]。因此,本文提出基于Gabor-SVM的高填方渠道水泥面破损检测模型研究,运用Gabor小波对高填方渠道的突变特征进行多尺度多方向特征提取,获得高填方渠道局部空间和频率域信息的良好特征,然后应用SVM进行破损分类和级别判断,充分利用SVM的小样本、非线性及高维模式识别的优势,以期得到更高的渗漏级别分类和识别率。
1 Gabor-SVM
1.1 Gabor小波
Gabor小波是一种纹理特征提取方法,2维Gabor函数描述的图像感受特性,与人类视觉系统中简单细胞的视觉刺激响应非常相似,在提取目标的局部空间和频率域信息方面具有良好特性[18-20]。2维Gabor的时域函数
$ g(x,y) = \left[ {\frac{1}{{2{\rm{ \mathsf{ π} }}{\sigma _x}{\sigma _y}}}} \right]\exp \left[ {\frac{1}{2}\left[ {\frac{{{x^2}}}{{\sigma _x^2\sigma _y^2}}} \right] + 2{\rm{ \mathsf{ π} }}jWx} \right] $ | (1) |
式中,可通过傅里叶变换到频域函数
$ G(u,v) = \exp \left\{ { - \frac{1}{2}\left[ {\frac{{{{(u - W)}^2}}}{{\sigma _u^2}} + \frac{{{v^2}}}{{\sigma _v^2}}} \right]} \right\} $ | (2) |
$ \left\{ {\begin{array}{*{20}{l}} {{\sigma _u} = \frac{1}{{2{\rm{ \mathsf{ π} }}{\sigma _x}}}}\\ {{\sigma _v} = \frac{1}{{2{\rm{ \mathsf{ π} }}{\sigma _y}}}} \end{array}} \right. $ | (3) |
式中,
$ {g_{mn}}\left( {{x^\prime },{y^\prime }} \right) = {a^{ - m}}g(x,y),a > 1,m,n \in \mathit{\boldsymbol{Z}} $ | (4) |
式中,
$ \left\{ {\begin{array}{*{20}{l}} {x' = {a^{ - m}}(x\cos \theta + y\sin \theta )}\\ {y' = {a^{ - m}}( - x\sin \theta + y\cos \theta )} \end{array}} \right. $ | (5) |
式中,
$ \left\{ {\begin{array}{*{20}{l}} {{u^\prime } = u\cos \theta + v\sin \theta }\\ {{v^\prime } = - u\sin \theta + v\cos \theta } \end{array}} \right. $ | (6) |
通过改变
1.2 SVM分类器
支持向量机(SVM)在解决小样本、非线性及高维模式识别中具有许多特有优势,并能够推广应用到函数拟合等其他机器学习问题中。此外,SVM最终解决的是一个凸二次规划问题,能够找到全局最优解[23-25]。SVM擅长通过松弛变量和核函数技术来解决数据线性不可分的问题。设数据样本集合为
$ wx + b = 0 $ | (7) |
式中,
若使满足式(7)的分类面对所有的数据样本都能正确分类,则须满足
$ {y_i}\left[ {w{x_i} + b} \right] - 1 \ge 0,i = 1, \cdots ,n $ | (8) |
式中,如果要实现
$ L(w,b,\alpha ) = \frac{1}{2}{\left\| w \right\|^2} - \sum\limits_{i = 1}^n {{\alpha _i}} \left[ {{y_i}\left( {w{x_i} + b} \right) - 1} \right] $ | (9) |
式中,
$ \left\{ {\begin{array}{*{20}{l}} {\frac{{\partial L}}{{\partial w}} = 0 \Rightarrow w = \sum\limits_{i = 1}^n {{\alpha _i}} {y_i}{x_i}}\\ {\frac{{\partial L}}{{\partial b}} = 0 \Rightarrow \sum\limits_{i = 1}^n {{\alpha _i}} {y_i} = 0}\\ {\frac{{\partial L}}{{\partial w}} \Rightarrow {\alpha _i}\left[ {{y_i}\left( {w{x_i} + b} \right) - 1} \right] = 0} \end{array}} \right. $ | (10) |
则最优分类面问题转化为求解在式(9)和式(10)约束条件下凸二次规划寻优的对偶问题,即
$ \begin{array}{*{20}{c}} {\max \sum\limits_{i = 1}^n {{\alpha _i}} - \frac{1}{2}\sum\limits_{i = 1}^n {\sum\limits_{j = 1}^n {{\alpha _i}} } {\alpha _j}{y_i}{y_j}\left( {{x_i},{x_j}} \right)}\\ {{\rm{ s}}{\rm{.t}}{\rm{. }}{\alpha _i} \ge 0,i = 1, \cdots ,n}\\ {\sum\limits_{i = 1}^n {{y_i}} {\alpha _i} = 0} \end{array} $ | (11) |
式(11)是一个存在唯一解的二次函数寻优的问题。设
$ {w^*} = \sum\limits_{i = 1}^n {\alpha _i^*} {y_i}{x_i} $ | (12) |
式中,
$ g(x) = sgn \left\{ {\sum\limits_{i = 1}^n {\alpha _i^*} {y_i}\left( {{x_i}x} \right) + {b^*}} \right\} $ | (13) |
式(13)中,
1.3 Gabor-SVM算法融合处理
图像经过Gabor滤波后一般包括幅值或相位等信息,由于Gabor相位信息不太稳定,会随着空间位置呈周期性变化。而幅值信息的变化相对稳定,并且反映了图像的能量谱。方差能反映图像高频部分的大小,图像对比度大,方差就大。均值能体现幅值的中心取值,起到一定区别作用。本文取Gabor幅值的方差与均值特征作为图像的Gabor表征。
Gabor-SVM的图像处理架构如图 1所示。整体框架首先对预处理好的样本图像
2 模型设计
3 实验过程及结果分析
3.1 破损等级分析
高填方渠道的破损类别可以定义为正常、裂缝、孔洞及破碎。裂缝相比图像背景亮度较暗,有一定的线性特征,如条状、锯齿条状等。孔洞是指水泥面板呈现孔或洞的破损现象,直径一般为2.5~10 mm,深度为10~50 mm,形状呈块状、圆形、椭圆形等的居多[28]。破碎由多条线性特征的裂缝组合而成,破损的缝比裂缝宽,是裂缝相互交叉。将坡面分割为3块以上的破损现象,复杂性大,尺度性明显,幅度的方差与均值均较高。实验中孔洞和裂缝同时出现的情况则判为破碎情况。若高填方渠道水泥坡面无以上3种破损情况,则称为正常,其幅度的方差与均值均较低。如图 3所示。
由于图像会受到各种噪声的干扰,因而在对图像进行特征提取处理前,一般会进行图像预处理,主要目的是消除图像中的无关信息,改进图像特征抽取的可靠性[29]。
3.2 特征分析
依照图 1处理框架,对多幅正常、裂缝、孔洞及破碎的图像进行Gabor滤波等处理,得到每个尺度/方向下的平均方差特征数据与平均均值特征数据,这些数据在尺度一定方向变化时,满足一定的大小关系,存在如下数值关系:孔洞>破碎>裂缝>正常。这种数值关系越明显,越有利于找出各种破损情况的线性关系。对高填方渠道的图像特征进行提取,得到当平均均值取7尺度、1~13方向时,满足孔洞>破碎>裂缝>正常的线性关系。平均方差在7尺度、连续1~13方向中也满足孔洞>破碎>裂缝>正常的线性关系。
由于4种类别的平均均值数值大小关系在多个尺度和方向下不明显,如在5尺度7~13方向时,其破碎与孔洞的数值差最小为0.01。若将无明显关系的数值添加到特征分类中,会使特征数据的冗余度增加,通过实验也验证了添加均值特征识别效果不理想。因此,本文仅取方差特征值进行分类识别。
若各类破损情况的方差差值越大,则其特征值越明显。于是对提取到的4个平均方差特征值分别进行两两相减求绝对值,取6个值中最小的差值(简称最小绝对差值)。最小绝对差值越大越平稳,各种情况的特征越明显,越有利于分类识别。通过对数据分析得到,最小绝对差值在1尺度7~13方向、2尺度5~13方向、5尺度9~12方向、6尺度9~13方向具有稳定值,分别约为0.46、1.40、0.56、0.88,差值最大的为2尺度的各方向,2尺度仅在方向取大于4的时候有平均方差满足一定的线性关系,部分方向的平均方差分布不明显。然而6尺度在多个方向下具有方差满足线性关系且最小绝对差值较大(0.88)。因此,取6尺度下的多个方向进行研究,找出特征最明显的尺度与方向参数组,用此参数组进行分类识别,验证特征提取分析的可靠性。
将6尺度在8~13方向未求平均方差的幅度方差在数轴上表示,如图 4所示。在这些尺度/方向参数组中,6尺度/12方向相对于其他参数组在特征值稳定性及取值范围方面更易于区分各类破损情况,幅度方差的分布特征明显,识别效果应最佳。因此,取6尺度12方向作为实验参数组。
3.3 结果分析
结合以上分析,本文共采集520幅图像,其中4类破损情况各90幅(共360幅)图像应用于Gabor-SVM算法训练、40幅(共160幅)图像进行模式识别测试。选择1~7尺度1~13方向的参数进行Gabor小波特征提取,得到的识别率数据用3维图表示,如图 5所示。
从图 5可以看出,正常的坡面图像识别率整体偏高,数据分布较为稳定,大部分分布在80%~100%之间。裂缝的坡面图像识别率从低到高稳定增长,但整体识别率低,大部分分布在50%~65%之间。孔洞的坡面图像识别率波动较大,与尺度的变化有明显关系,如1、2尺度时识别率低,3、4、5尺度时识别率逐渐增高,6、7尺度时识别率降低,识别率大部分分布在78%~88%之间。破碎的坡面图像识别率增长与尺度的大小有一定关系:从低到高再到低,波动较大,识别率大部分在80%~95%之间。通过识别率大小比较发现,均值为85%的有4/9、5/11、6/12、6/13(尺度/方向)参数组,6尺度/12方向时特征的提取效果在最佳范围内,这与3.2节特征的分析结论相对应,因此取6尺度/12方向为最佳识别的参数组。在6尺度/12方向时,孔洞、破碎、裂缝、正常的识别率分别为:88%、90%、63%、98%。
3.4 各种特征提取方法比较
在相同环境下,进行了直方图-SVM、灰度共生矩阵-SVM、Canny边缘检测算法-SVM、和Gabor-欧氏距离算法的实验,并与Gabor-SVM的识别效果进行误差分析,如表 1和表 2所示。从表 1和表 2可以看出,直方图-SVM、灰度共生矩阵-SVM、Canny算法-SVM和Gabor-欧氏距离算法对正常的坡面图像识别率整体偏高,对裂缝和孔洞的坡面图像识别率偏低,与其他方法相比,Gabor-SVM的裂缝识别效果整体较好。
表 1
多种算法处理的识别率
Table 1
Recognition rates of multiple algorithms
/% | ||||
方法 | 破损类别 | |||
正常 | 裂缝 | 孔洞 | 破碎 | |
直方图-SVM | 98 | 0 | 25 | 65 |
灰度共生矩阵-SVM | 55 | 25 | 13 | 68 |
Canny算法-SVM | 88 | 30 | 0 | 75 |
Gabor-欧氏距离 | 73 | 15 | 25 | 90 |
Gabor-SVM | 98 | 63 | 88 | 90 |
表 2
识别率不小于70%的判别表
Table 2
Discriminant table with recognition rate not less than 70%
方法 | 破损类别 | |||
正常 | 裂缝 | 孔洞 | 破碎 | |
直方图-SVM | √ | × | × | × |
灰度共生矩阵-SVM | × | × | × | × |
Canny算法-SVM | √ | × | × | √ |
Gabor-欧氏距离 | √ | × | × | √ |
Gabor-SVM | √ | × | √ | √ |
表 1中,从Gabor-SVM与Gabor-欧氏距离两种方法来看,Gabor-欧氏距离的正常及破碎的坡面图像情况分类匹配效果较好,但Gabor-欧氏距离的裂缝与孔洞坡面图像识别率较低,体现出其识别和分类的不稳定性。从表 2可以看出,Gabor-SVM模型对水泥坡面的孔洞、正常及破碎等识别率较高,但对裂缝的识别效果不理想。其原因是裂缝的宽度与背景图片对比度低,在进行多方向/尺度划分后,会将裂缝的条纹细化,从而识别错的概率增大。
4 结论
本文设计了基于Gabor-SVM的南水北调高填方渠道水泥坡面破损监测模型,在相同环境下进行了直方图-SVM、灰度共生矩阵-SVM、Canny算法-SVM及Gabor-欧氏距离的破损识别实验和对比分析。结论如下:
1) 通过Gabor小波提取的多尺度/方向特征对比分析,得出在6尺度/12方向时各种破损等级的特征明显,识别效果最佳。
2) Gabor-SVM算法模型对渠坡正常情况的识别效果最高,但对裂缝情况的识别效果不理想,平均识别率为85%。
3) 整体来说,基于Gabor-SVM的图像检测法实现了高填方渠坡破损监测的突变特征识别和等级划分的功能,可为查找高填方渠道渗漏隐患提供技术支持。
在实际的高填方渠道渗漏检测工作中,由于采集数据的环境变化问题,渠道坡面图像的多尺度/方向的特征会随环境变化而变化,本文提出的Gabor-SVM算法模型还需要重新进行训练与测试。另外,文中采集的图像数据来源于迎水坡水面以上的图像,还没有进行水下的坡面图像处理。下一步的工作将会在多种环境下进行高填方渠道坡面的破损识别,并提高Gabor-SVM算法模型的识别率和可靠性,以扩大该方法的使用范围。
参考文献
-
[1] Cui G, Chen J S, Wang L L. Design scheme of seepage flow monitoring in high fill section of south-to-north water diversion project[J]. West-China Exploration Engineering, 2013, 25(2): 39–41. [崔岗, 陈俊生, 王丽丽. 南水北调高填方段渗流监测设计方案[J]. 西部探矿工程, 2013, 25(2): 39–41. ] [DOI:10.3969/j.issn.1004-5716.2013.02.014]
-
[2] Qu Z G, Shen L P, Li M X, et al. Effective reinforcement measures for high-filled canal of middle route project of south-to-north water diversion[J]. Yangtze River, 2013, 44(16): 63–66. [屈志刚, 申黎平, 李明新, 等. 南水北调中线工程高填方渠道加强措施探讨[J]. 人民长江, 2013, 44(16): 63–66. ] [DOI:10.3969/j.issn.1001-4179.2013.16.018]
-
[3] Guo H, Chen X D, Liu G. Study on the practices of water rights trading in south-to-north water diversion middle route project[J]. South-to-North Water Transfers and Water Science & Technology, 2018, 16(3): 175–182. [郭晖, 陈向东, 刘钢. 南水北调中线工程水权交易实践探析[J]. 南水北调与水利科技, 2018, 16(3): 175–182. ] [DOI:10.13476/j.cnki.nsbdqk.2018.0085]
-
[4] Hou Q W, Feng X B, Li Z J, et al. Physical simulation-based experimental study on acting force from ice-cover on canal of the south-to-north water diversion middle route project[J]. Water Resources and Hydropower Engineering, 2018, 49(2): 63–69. [侯倩文, 冯晓波, 李志军, 等. 冰盖对南水北调中线工程渠道作用力物理模拟试验研究[J]. 水利水电技术, 2018, 49(2): 63–69. ] [DOI:10.13928/j.cnki.wrahe.2018.02.010]
-
[5] Coggon J H. Electromagnetic and electrical modeling by the finite element method[J]. Geophysics, 1971, 36(1): 132–155. [DOI:10.1190/1.1440151]
-
[6] Lin C H, Lin C P, Hung Y C, et al. Application of geophysical methods in a dam project:life cycle perspective and Taiwan experience[J]. Journal of Applied Geophysics, 2018, 158: 82–92. [DOI:10.1016/j.jappgeo.2018.07.012]
-
[7] Bernstone C, Heyden A. Image analysis for monitoring of crack growth in hydropower concrete structures[J]. Measurement, 2009, 42(6): 878–893. [DOI:10.1016/j.measurement.2009.01.007]
-
[8] Fu X H. Seepage analysis of south to north water transfer midline high fill canal segment[D]. Zhengzhou: North China University of Water Resources and Electric Power, 2014. [付新航.南水北调中线干线工程高填方渠段渗流分析[D].郑州: 华北水利水电大学, 2014. [DOI: 10.7666/d.D543214]]
-
[9] Tian Z Z. Research on the multi-source data fusion model for leakage monitoring of high filled channel of south-to-north water transfer project[D]. Zhengzhou: North China University of Water Resources and Electric Power, 2017. [田壮壮.南水北调高填方渠道渗漏监测的多源数据融合模型研究[D].郑州: 华北水利水电大学, 2017.]
-
[10] Yang X, Yang X. An algorithm for estimating the precision of scene guidance based on Gabor wavelet[J]. Computer Engineering, 2003, 29(18): 72–73, 103. [杨昕, 杨新. 一种基于Gabor小波的制导精度预测方法[J]. 计算机工程, 2003, 29(18): 72–73, 103. ] [DOI:10.3969/j.issn.1000-3428.2003.18.030]
-
[11] Yang J H, Yao J P, Fan L, et al. Research on veneer surface crack image processing methods[J]. Forestry Machinery & Woodworking Equipment, 2017, 45(4): 29–31. [杨建华, 姚建平, 范磊, 等. 单板表面裂缝图像处理方法研究[J]. 林业机械与木工设备, 2017, 45(4): 29–31. ]
-
[12] Sun L, Xing J C, Xie L Q, et al. An adaptive threshold-based canny algorithm for crack detection[J]. Microcomputer & Its Applications, 2017, 36(5): 35–37, 41. [孙亮, 邢建春, 谢立强, 等. 基于自适应阈值Canny算法的裂缝检测方法研究[J]. 微型机与应用, 2017, 36(5): 35–37, 41. ] [DOI:10.19358/j.issn.1674-7720.2017.05.012]
-
[13] Cheng C H. The analysis of concrete cracks'fractal and graylevel cooccurrence matrix[J]. Low Temperature Architecture Technology, 2014, 36(7): 13–15. [程昌宏. 混凝土裂缝分形和灰度共生矩阵分析[J]. 低温建筑技术, 2014, 36(7): 13–15. ] [DOI:10.3969/j.issn.1001-6864.2014.07.005]
-
[14] Hadizadeh H. Multi-resolution local Gabor wavelets binary patterns for gray-scale texture description[J]. Pattern Recognition Letters, 2015, 65: 163–169. [DOI:10.1016/j.patrec.2015.07.038]
-
[15] Xu W. Research of face recognition methods based on Gabor features and SVM[D]. Urumqi: Xinjiang University, 2015. [许伟.基于Gabor特征和SVM的人脸识别方法研究[D].乌鲁木齐: 新疆大学, 2015.]
-
[16] Fisher W D, Camp T K, Krzhizhanovskaya V V. Crack detection in earth dam and levee passive seismic data using support vector machines[J]. Procedia Computer Science, 2016, 80: 577–586. [DOI:10.1016/j.procs.2016.05.339]
-
[17] Xia M, Lu W T, Yang J, et al. A hybrid method based on extreme learning machine and $k$-nearest neighbor for cloud classification of ground-based visible cloud image[J]. Neurocomputing, 2015, 160: 238–249. [DOI:10.1016/j.neucom.2015.02.022]
-
[18] Xing Y, Yang Q S, Guo C A. Face recognition based on Gabor enhanced marginal fisher model and error correction SVM[C]//Proceedings of the 8th International Symposium on Neural Networks. Guilin, China: Springer-Verlag, 2011: 296-304.[DOI: 10.1007/978-3-642-21090-7_35]
-
[19] Zhang G, Ma Z M. An approach of using Gabor wavelets for texture feature extraction[J]. Journal of Image and Graphics, 2010, 15(2): 247–254. [张刚, 马宗民. 一种采用Gabor小波的纹理特征提取方法[J]. 中国图象图形学报, 2010, 15(2): 247–254. ] [DOI:10.11834/jig.20100210]
-
[20] Chaudhari S, Gulati R M. Script identification using Gabor feature and SVM classifier[J]. Procedia Computer Science, 2016, 79: 85–92. [DOI:10.1016/j.procs.2016.03.012]
-
[21] Qi D F, Wu C D. Face recognition based on Gabor wavelet transform and support vector machine[J]. Science & Technology Information, 2009(35): 31–32. [戚大方, 吴成东. 基于Gabor小波变换与支持向量机的人脸识别[J]. 科技信息, 2009(35): 31–32. ] [DOI:10.3969/j.issn.1001-9960.2009.35.344]
-
[22] Li Q X. Research on face expression recognition algorithm based on improved Gabor wavelet feature extraction[D]. Qinhuangdao: Yanshan University, 2017. [李清霞.基于改进的Gabor小波特征提取的人脸表情识别算法研究[D].秦皇岛: 燕山大学, 2017.]
-
[23] Panja R, Pal N R. MS-SVM:minimally spanned support vector machine[J]. Applied Soft Computing, 2018, 64: 356–365. [DOI:10.1016/j.asoc.2017.12.017]
-
[24] Liu Y, Wen K W, Gao Q X, et al. SVM based multi-label learning with missing labels for image annotation[J]. Pattern Recognition, 2018, 78: 307–317. [DOI:10.1016/j.patcog.2018.01.022]
-
[25] Shi Y G, Cheng K, Liu Z W. Segmentation of hippocampal subfields by using deep learning and support vector machine[J]. Journal of Image and Graphics, 2018, 23(4): 542–551. [时永刚, 程坤, 刘志文. 结合深度学习和支持向量机的海马子区图像分割[J]. 中国图象图形学报, 2018, 23(4): 542–551. ] [DOI:10.11834/jig.170431]
-
[26] Zhang Y, Chen S X, Yu F, et al. Calculation method study of settlement process of high filling channels in south-to-north water diversion project[J]. Chinese Journal of Rock Mechanics and Engineering, 2014, 33(S2): 4367–4374. [张宇, 陈善雄, 余飞, 等. 南水北调高填方渠道沉降过程计算方法研究[J]. 岩石力学与工程学报, 2014, 33(S2): 4367–4374. ] [DOI:10.13722/j.cnki.jrme.2014.s2.122]
-
[27] Xin H Y. Construction technology of plastic concrete barrier wall in high fill canal section[J]. Henan Water Resources & South-to-North Water Diversion, 2012(18): 134–135. [辛贺艳. 高填方渠段塑性混凝土防渗墙施工技术[J]. 河南水利与南水北调, 2012(18): 134–135. ]
-
[28] Wang Y X. On reasons for damages of the cement concrete pavement[J]. Shanxi Architecture, 2009, 35(5): 275–276. [王彦须. 水泥混凝土路面破损原因分析[J]. 山西建筑, 2009, 35(5): 275–276. ] [DOI:10.3969/j.issn.1009-6825.2009.05.174]
-
[29] Luo R. Research of pavement crack detection algorithm based on Image processing[D]. Wuhu: Anhui Polytechnic University, 2017. [罗瑞.基于图像处理的路面裂缝检测算法研究[D].芜湖: 安徽工程大学, 2017.]