English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

As underwater environments are complex and time-varying, and the target is randomness and diversity. Underwater dam crack detection based on autonomous underwater vehicle (AUV) equipped with sonar is a challenging task. In this paper, a novel crack block tree (BT) algorithm utilized in sonar images is proposed to detect underwater dam crack. Firstly the low-resolution sonar images should be pre-processing, then the images are divided into image blocks for clustering analysis. After that, an adaptive fracture fragments based on tensor voting is put forwarded to combine the crack segment. At last, a minimum spanning tree is used to obtain the cracks. By comparing the crack BT with other algorithms, the results show that the proposed approach can deal with the crack detection more effectively. Introduction Dam crack detection is an important work to ensure the dam safety [1] The dam crack is a mutation in the process of its internal damage accumulation [2]. Thus, it is always used to indicate the danger degree of the dam damage which is attracted by many scholar's attention. Recently, the underwater dam crack detection with sonar images has become one of the most important methods for its nondestructive, intuitive, convenient and efficient [3]. However, underwater environments are complex, time-varying and full of interference [4-6] which will lead to the measurement signal submerged by noise. Besides, the unstructured crack is random and diverse leading to the difficulties to describe the cracks. Aimed at those problems mentioned above, various methods have been proposed. Chen, et al. [7] presented an adaptive underwater dam crack edge detection algorithm based on multi-structures and multi-scale elements. Bernstone, et al. [8] proposed a digital image analysis technique for crack monitoring using a standard web-camera to acquire continuous data sets in concrete dams. Xu, et al. [9] suggested an integrated model used digital image processing in developing the numerical representation of concrete structures defects. However, most of the traditional detection are proposed by exploiting characteristics which always submerged by noise. It will lead to low detection rate and high false alarm rate. In this paper, a novel crack detection algorithm called crack BT is proposed. Section 2 illustrates the basic idea of the crack BT. Section 3 gives the result of proposed algorithm and analysis. The crack BT Before the crack BT, some prior knowledge about crack can be introduced: 1) the image regions are darker than their surroundings; 2) the connected domain of the crack region is slender than other regions. The workflow chart of the proposed approach is shown in Fig. 1, which will be introduced in details as follows. Sonar image K-means clustering Tensor voting Minimum spanning tree The result Fig.1. The workflow of the crack BT 5th International Conference on Computer Sciences and Automation Engineering (ICCSAE 2015) © 2016. The authors Published by Atlantis Press 452 Clustering analysis of the image blocks. The sonar images should be pre-processing first, and the details can be seen in previous work [9]. After that, the image is divided into nonoverlapping image blocks which is shown in Fig. 2(b). Three characteristic matrices are computed for each block: 1) the mean value matrix (Mm); 2) the standard deviation matrix (STDm); 3) the information entropy matrix (IFE). A clustering analysis method is used to distinguish crack (crack-like) from background with the 3-D space classifiers, and each point identifies one image block. Then the feature set in the 3-D space is defined as: 3 1 1 1 2 {( , ) ( , ) : ; { , }} n n i j F V y V y V R y c c = ∈ ∈  (1) where n is the number of block points for the pattern vector V, which is constructed by Mm, STDm and IFE; i y corresponds to the i th block. The blocks are divided into two classes, 1 c and 2 c , denoting blocks carrying non-crack information and the blocks containing with cracks information respectively. Furthermore, the K-means classification approach is used to make the clustering analysis [10]. The clustering result in the 3-D feature space is shown in Fig. 2(i). The classes that are labeled with red circle in the 3-D feature space are the target class 2 c . The final results of the target blocks are shown in Fig. 2(c), which are marked with red squares. Adaptive tensor voting of the crack fragments. Cracks marked with independent segments after clustering analysis are actually an integrated crack. It will affect the understanding of the crack if the segments are not connected. The spatial proximity and the gliding property of the crack fragments is the main difference with the interference. Thus, a self-adaptive tensor voting algorithm is put forwarded. The images can be expressed in a tensor field which contains ball tensor and stick tensor. The two tensors can be defined as follow: Ball tensor: if point P is an isolated point, the tensor is expressed as 1 0 0 1       . Stick tensor: if P is the point on the curve, the tensor is expressed as 2 2 cos sin co sin n s s cos i θ θ θ θ θ θ       . θ is the tangent angle between the tangent and the horizontal direction. The stick tensor along the y-axis from the coordinate origin O, and the voting strength from P can be calculated by degradation function [11]:

A novel underwater dam crack detection algorithm based on sonar images