scholarly journals Microwave Nondestructive Testing for Defect Detection in Composites Based on K-means Clustering Algorithm

IEEE Access ◽  
2020 ◽  
pp. 1-1
Author(s):  
Nawaf H. M. M. Shrifan ◽  
Ghassan N. Jawad ◽  
Nor Ashidi Mat Isa ◽  
Muhammad Firdaus Akbar
Keyword(s):  
Nondestructive Testing ◽  
Defect Detection ◽  
Clustering Algorithm ◽  
Microwave Nondestructive Testing

Метод и аппаратура инфракрасного и ультразвукового термографического контроля крупногабаритных композиционных изделий сложной формы

2021 ◽  
Vol 7 ◽  
pp. 67-74
Author(s):  
А.О. Чулков ◽  
Д.А. Нестерук ◽  
Б.И. Шагдыров ◽  
В.П. Вавилов
Keyword(s):  
Neural Network ◽  
Nondestructive Testing ◽  
Data Fusion ◽  
Defect Detection ◽  
Complex Shape ◽  
Reference Sample ◽  
Robotic System ◽  
Image Stitching ◽  
Large Size ◽  
Thermal Nondestructive Testing

A robotic system for combined thermal nondestructive testing of large-size parts, including data fusion, is described. The efficiency of combining results of infrared (IR) and ultrasonic IR thermographic inspection has been demonstrated on a complex-shape reference sample containing 18 surrogates of manufacture and in-service defects. The data fusion algorithms including IR image stitching in space and automated defect detection and characterization by using a neural network have demonstrated efficiency of the proposed approach in practical testing.

scholarly journals Wood Defect Detection Based on Depth Extreme Learning Machine

2020 ◽  
Vol 10 (21) ◽  
pp. 7488
Author(s):  
Yutu Yang ◽  
Xiaolin Zhou ◽  
Ying Liu ◽  
Zhongkang Hu ◽  
Fenglong Ding
Keyword(s):  
Deep Learning ◽  
Extreme Learning Machine ◽  
Defect Detection ◽  
Clustering Algorithm ◽  
Classification Performance ◽  
Image Features ◽  
Test Time ◽  
Machine Model ◽  
Feature Extraction Method ◽  
Learning Machine

The deep learning feature extraction method and extreme learning machine (ELM) classification method are combined to establish a depth extreme learning machine model for wood image defect detection. The convolution neural network (CNN) algorithm alone tends to provide inaccurate defect locations, incomplete defect contour and boundary information, and inaccurate recognition of defect types. The nonsubsampled shearlet transform (NSST) is used here to preprocess the wood images, which reduces the complexity and computation of the image processing. CNN is then applied to manage the deep algorithm design of the wood images. The simple linear iterative clustering algorithm is used to improve the initial model; the obtained image features are used as ELM classification inputs. ELM has faster training speed and stronger generalization ability than other similar neural networks, but the random selection of input weights and thresholds degrades the classification accuracy. A genetic algorithm is used here to optimize the initial parameters of the ELM to stabilize the network classification performance. The depth extreme learning machine can extract high-level abstract information from the data, does not require iterative adjustment of the network weights, has high calculation efficiency, and allows CNN to effectively extract the wood defect contour. The distributed input data feature is automatically expressed in layer form by deep learning pre-training. The wood defect recognition accuracy reached 96.72% in a test time of only 187 ms.

An Infrared Thermal Image Processing Framework for Defect Detection of a Metal Part with Rough Surface

2012 ◽  
Vol 229-231 ◽  
pp. 1356-1360
Author(s):  
Jing Xie ◽  
Chang Hang Xu ◽  
Guo Ming Chen
Keyword(s):  
Image Processing ◽  
Rough Surface ◽  
Similarity Measure ◽  
Defect Detection ◽  
Clustering Algorithm ◽  
Thermal Image ◽  
Metal Part ◽  
Fuzzy C Means ◽  
Fuzzy C Means Clustering ◽  
Processing Framework

We propose an infrared thermal image processing framework based on a modified fuzzy c-means clustering algorithm with revised similarity measure in this paper. The framework can realize the defect detection of a metal part with rough surface. Firstly, a comprehensive method is used to preprocess infrared thermal image. Secondly, the preprocessed image is segmented using modified fuzzy c-means clustering algorithm with revised similarity measure. Finally, taking the average gray level of each cluster in the original gray scale image as a feature, defect cluster is recognized. Experimental result shows that the proposed framework has very promising performance and can obtain precise information of defects on a metal part with rough surface.

Sign in / Sign up

Export Citation Format

Share Document