Automatic Defect Recognition Research Articles

Image Based Positioning of the Freightcar’s Running Gear Springs

Freightcar’s running gear springs are easily ruptured in the processes of manufacturing and using, losing the role of relaxing vehicle vibration, and resulting in accident. It is time to study the automatic recognition of springs defects based on image processing technology. Defects recognition is on the premise of accurate spring image location. An image location algorithm for the freightcar’s running gear springs is proposed in this paper, which is based on the structure characteristics of springs themselves. The algorithm mainly includes graying the image, correcting non-uniform illumination by wavelet homomographic filtering, taking binaryzation on image based on the iterative threshold method and locating springs’ boundary location, which could be obtained by image scanning column by column and feature analysis. This algorithm can achieve the automatic positioning of springs and has a good result to locate springs in the running gear images in case of different illumination; the accuracy rate has reached 96 percent. The study of this paper lays a foundation on the realization of automatic recognition of springs defects and has good application prospects.

Automatic Detection of Porosity and Slag Inclusion in Boilers using Statistical Pattern Recognition Techniques

automatic weld inspection system can be used for training weld quality inspectors and for weld quality assessment in the welding industry. The system is aimed at increasing inspection speed, accuracy and at the same time reducing subjectivity associated with the manual interpretation of the weld radiographs. The system detects the weld defects, classifies it, and displays the output with greater accuracy compared to the previously existing methods. It implements fuzzy logic that helps in the automatic recognition of weld defects efficiently.

Visual Pattern Recognition Supporting Defect Reporting and Condition Assessment of Wastewater Collection Systems

Computational approaches to visual data and information processing can assist asset management procedures. In this paper, we present a computerized visual pattern recognition approach to facilitate inspection and condition assessment of wastewater collection systems in realistic settings. This research aims to provide a basis for condition assessment and defect reporting in wastewater infrastructure, enabling automatic detection and recognition of defects and patterns from inspection images or videos. An approach for sensing and automatically detecting critical areas or regions of interest, and further discovering and recognizing objects of interest in large image data sets, is described. The major steps of the approach are illustrated by examples using actual inspection images acquired from wastewater pipelines.

GenLab: A MATLAB ®-based program for structural analysis of folds mapped by GPS or seismic methods

GenLab is a MATLAB ® code for structural analysis of digitally and seismically mapped horizons based on the geometrical model of the developable surface. This model is chosen because surfaces of this type result from the bending of a non-stretching stratum and because deviation from developable geometry implies that bed stretching during folding has occurred. The model is therefore linked with straining of the horizon and can be used for highlighting regions of deformation expressed by brittle fracturing or ductile deformation. Mathematically, a developable surface is one upon which straight lines (generators) can be drawn, and along these generators the orientations of tangent planes of the surface remain constant. Our method involves treating real geological surfaces as approximations to developable surfaces and determines the generator lines at different points on the surface. At these points we consider strip samples of the surface radiating in different directions and examine the variability of strike and dip within each sample strip. The radial scan-line length parameter is user-defined, and the optimal value can be easily identified by geostatistical analysis of the depth data. The local plunge lines (the approximation to the generatrix of a developable fold) are therefore identified as the direction associated with the least variation of surface attitude, i.e., the most constant strike and dip. The map pattern obtained by analyzing the plunges and trends of plunge lines across an area allows identification of domains where folding accords closely with a developable geometry. Such domains are recognized from straight plunge lines, defining parallel or convergent patterns. According to the model, deviations from these patterns correspond to regions of structural complexity associated with ductile or brittle straining of the surface being analyzed. The plunge-line analysis therefore offers a useful technique with potential for predicting strained zones and automatic fault recognition. GenLab allows the rapid implementation of this analysis and uses the matrix capabilities and contouring functions of MATLAB.

High-speed CT imaging for today's industry

Computed Tomography (CT) has been extensively used in industrial inspection, however, production applications have been few, due to the time required for performing CT. This paper describes a high-throughput, production X-ray CT system for inspecting munitions. Various types of defects such as cracks, voids, fuse armed status, and powder fill level are identified and evaluated. Automatic defect recognition algorithms and software have been developed to inspect the CT data, due to the large amount of data involved, and high throughput required.

A model-based clustering approach to the recognition of the spatial defect patterns produced during semiconductor fabrication

Defects on semiconductor wafers tend to cluster and the spatial defect patterns of these defect clusters contain valuable information about potential problems in the manufacturing processes. This study proposes a model-based clustering algorithm for automatic spatial defect recognition on semiconductor wafers. A mixture model is proposed to model the distributions of defects on wafer surfaces. The proposed algorithm can find the number of defect clusters and identify the pattern of each cluster automatically. It is capable of detecting defect clusters with linear patterns, curvilinear patterns and ellipsoidal patterns. Promising results have been obtained from simulation studies.

Pattern recognition based digital relaying for advanced series compensated line

This paper presents a new approach for fault analysis of advanced series compensated (thyristor controlled series compensated) line using pattern recognition approach. Here, S-transform (ST) is used to process the post fault current signal samples and features are extracted to identify the faulty phase and the faulty section involved in the fault process. The S-transform is an extension of wavelet transform, which possesses superior property over the latter as the moving functions are fixed with respect to time axis while the localizing scalable Gaussian window dilates and translates. Such a transform with moving and scalable localizing Gaussian window, therefore, provides excellent time localization property for different signals. Utilizing the ST, the time-frequency contours (ST coefficients) are generated for various fault conditions in advanced series compensated transmission system. Also an automatic fault recognition system is designed using probabilistic neural network (PNN) to provide the faulty phase and faulty section involved in the TCSC based line. The PNN is tested and trained with the features extracted from the ST contours for different fault situations. The proposed method is tested for all 11 types of shunt faults with wide range of operating conditions and provides accurate results.

Image Processing Algorithm for Quantitative Characterization of Thermal Imaging Acquired During On-line Weld Monitoring

Defect in welded structures is a matter of serious concern. The current practice involves interpretation by inspectors and experts, which is time consuming. With greater emphasis on automation during manufacturing process, automated NonDestructive Testing (NDT) Techniques have gained prominence. The key to success of automated NDT lies in the automatic defect recognition and characterization. Dimensional Characterization of defects or discontinuities is essential in order to compare this with the acceptable codes and standards. Greater emphasis is on selecting appropriate mathematical tools so that feature extraction is possible in accurate and reliable manner. Conventionally image-processing algorithms are applied to extract and quantify the features depicting defects. However many image processing algorithms are available for specific applications with its advantages and disadvantages. This paper explores the possibility of image processing algorithms like edge detection and morphological operators for feature extraction and proposes an algorithm for quantitative measurements of defects on thermal images. Lack of Fusion and Tungsten Inclusion are the defects considered. The choice of these particular defects is that these defects are the main causes for rejection and can be detected reliably during thermal imaging.

Structural analysis of seismically mapped horizons using the developable surface model

A new method for the analysis of folding of seismically mapped horizons is described. Based on a model of developable surfaces, the local geometrical properties are determined by analyzing the variation of dip and strike along linear strips on the surface. By considering strips in different directions, a plunge line (the approximation to the generatrix of a developable fold) is identified as the direction associated with the least variation of surface attitude. The map pattern obtained by analyzing the plunge and trend of plunge lines across an area allows the identification of domains where folding accords with a developable geometry. Such domains are recognized from straight plunge lines, defining convergent or parallel patterns. Deviations from these patterns correspond to regions of structural complexity associated with ductile or brittle straining of the horizon. We suggest that plunge-line analysis may offer a useful technique for automatic fault recognition.

터보회전기기의 진동모니터링 및 진단을 위한 이산 은닉 마르코프 모델에 관한 연구

Condition monitoring is very important in turbo machinery because single failure could cause critical damages to its plant. So, automatic fault recognition has been one of the main research topics in condition monitoring area. We have used a relatively new fault recognition method, Hidden Markov Model(HMM), for mechanical system. It has been widely used in speech recognition, however, its application to fault recognition of mechanical signal has been very limited despite its good potential. In this paper, discrete HMM(DHMM) was used to recognize the faults of rotor system to study its fault recognition ability. We set up a rotor kit under unbalance and oil whirl conditions and sampled vibration signals of two failure conditions. DHMMs of each failure condition were trained using sampled signals. Next, we changed the setup and the rotating speed of the rotor kit. We sampled vibration signals and each DHMM was applied to these sampled data. It was found that DHMMs trained by data of one rotating speed have shown good fault recognition ability in spite of lack of training data, but DHMMs trained by data of four different rotating speeds have shown better robustness.

Condition assessment of underground sewer pipes using a modified digital image processing paradigm

The National Science Foundation (NSF) has estimated the total US investment in civil infrastructure at $US 20 trillion. The investments in sewer pipeline collection systems represent a major component of this figure. Recent studies have shown that to upgrade or restore municipal infrastructure in North America to an acceptable level the cost will be in excess of $US 100 billion. Many of these systems are eroding due to aging, excessive demand, misuse, poor construction, mismanagement, and neglect. Visual inspection based on Closed Circuit Television (CCTV) surveys is used widely in North America to assess the condition of underground sewer pipes. The human eye is extremely effective at recognition and classification, but it is not suitable for assessing pipe defects in thousand of miles of pipeline due to fatigue, subjectivity, and cost. Automatic recognition of various pipe defects is of considerable interest since it has the potential to overcome problems of fatigue, subjectivity, and ambiguity, leading to economic benefits. In this paper, we present a system for the application of computer vision techniques to the automatic assessment of the structural condition of underground sewer pipes that overcomes the inherent limitations of existing digitizing paradigms. An image enhancement method that eliminates the non-uniform background illumination while ensuring that the enhanced image will be faithful to the source image and enables quantification of distress features is presented. Algorithms for analysis of joint and lateral recognition are also presented and should allow engineers to be able to gain the goal of a fully automated underground sewer defect detection system based on digital image processing of SSET surveys according to the principles demonstrated in this paper.

Automatic recognition of weld defects in x-ray inspection : Daum, W.; Rose, P.; Heidi, H.; Builtjes, J.H. British Journal of Nondestructive Testing, Vol. 29, No. 2, pp. 79–81 (Mat. 1987)

Automatic recognition of weld defects in x-ray inspection : Daum, W.; Rose, P.; Heidi, H.; Builtjes, J.H. British Journal of Nondestructive Testing, Vol. 29, No. 2, pp. 79–81 (Mat. 1987)

Neural network classification of flaws detected by ultrasonic means

A methodology for the automatic recognition of weld defects, detected by a P-scan ultrasonic system, has been developed within two stages in the present work. In the first stage, a selection of the shape parameters defining the pulse-echo envelope reflected from a generic flaw, and defined in the time domain, is performed by Fischer linear discriminant analysis. In the second stage the classification is carried out by a three-layered neural network trained with the backpropagation rule, where the input values are the parameters selected by the Fischer analysis. With regard to the neural network learning process, 135 real weld defects have been considered. The defects, distributed among the classes of cracks, slags of inclusion and porosity, had been previously characterized by X-ray inspection. The results obtained confirm the effectiveness of the approach in preserving the discriminant information needed for characterization by an iterative use of Fischer analysis, and in increasing the generalization properties of the layered network by an interpretation of the knowledge embedded in the generated connections and weights. The required computation time allows in-process application.

Automatic recognition of weld defects in x-ray inspection : Daum, W.; Rose, P.; Heidt, H.; Builtjes, J.H. British Journal of Nondestructive Testing, Vol. 29, No. 2, pp. 79–81 (Mat. 1987)

Automatic recognition of weld defects in x-ray inspection : Daum, W.; Rose, P.; Heidt, H.; Builtjes, J.H. British Journal of Nondestructive Testing, Vol. 29, No. 2, pp. 79–81 (Mat. 1987)

A knowledge-based image-inspection system for automatic defect recognition, classification, and process diagnosis

Combining knowledge-based processing with image processing is a key issue in the future of the visual inspection of complex patterns such as offset prints. Often the class of the defect determines the state of the process, which must known for eliminating the cause of the defect. We describe the architecture of such a complex knowledge-based inspection system. The system has been used for defect recognition and misprint diagnosis in offset printing, but it is flexible enough for other applications. The system is based on a set of general and powerful tools for the knowledge interpretation of sensor signals. An object-oriented concept and task-dependent algorithms for efficient image processing are implemented. The paper concentrates on four points: integration of the system in the offset printing process, a description of the system architecture, knowledge acquisition, and implementation results.

Automatic defects recognition in composite aerospace structures from experimental and theoretical analysis as part of an intelligent infrared thermographic inspection system : David, D.; Marin, J.Y.; Tretout, H. Thermosense XIV. An International Conference on Thermal Sensing and Imaging Diagnostics Applications, Orlando, Florida (United States), 22–24 Apr. 1992, pp. 182–193. SPIE Vol. 1682 (1992)

Automatic defects recognition in composite aerospace structures from experimental and theoretical analysis as part of an intelligent infrared thermographic inspection system : David, D.; Marin, J.Y.; Tretout, H. Thermosense XIV. An International Conference on Thermal Sensing and Imaging Diagnostics Applications, Orlando, Florida (United States), 22–24 Apr. 1992, pp. 182–193. SPIE Vol. 1682 (1992)

Automatic fault recognition by image correlation neural network techniques

The image correlation technique, a useful method for online inspection for quality production control, is discussed. A Cauchy machine determines the imperfection by the degree of orthogonality between the automated extracted feature from the send-through image and the class feature of early good samples. The performance measure used for such an automatic feature extraction is based on a certain minimax cost function useful for image classification. Such an inspection theory based on image sequences is simulated by incorporating space-filling Peano curves, fast simulated Cauchy annealing, and minimax classification performance measures. An artificial neural network (ANN) is discussed as a possible implementation. >

Automatic recognition of weld defects in x-ray inspection

Automatic recognition of weld defects in x-ray inspection

Automatic recognition of weld defects in x-ray inspection

Automatic recognition of weld defects in x-ray inspection

Computer aided ammunition radiography : Wu, K.K.M.; Quinn, J.D.; Hedger, D.F. Materials research laboratories, Ascot Vale, Victoria (Australia), AD-A160-759, vp (Jul. 1985)

Abstract : A need exists for improving the current methods for ensuring the safety of ammunition fuzes. Accordingly the feasibility of using automatic techniques for the inspection of ammunition fuzes has been investigated. The investigation covered the following aspects. (1) The usefulness of image enhancement techniques for the improvement of the detectability of defects; (2) the feasibility of automatic recognition of defects: and (3) the merits and the practicability of using real-time radiography to replace the film-based system for fuze inspection. It may be concluded from the feasibility study that image qualities can be significantly enhanced for visual detection. When a defect can be characterised in terms of the intensity and geometrical factors of its radiographic images automatic identification of the defect also appears to be feasible. The assessment also indicated that the use of real-time radiography is now a practical proposition. Keywords: Pattern recognition; Image processing; Australia. (Author)