Qiang Song

A New Multi-scale Texture Analysis with Structural Texel

This paper presents a new methodology for multi-scale texture analysis. The basic idea is that an image texture is viewed as a tessellation of square texels of different sizes and pixel levels. A textural image is decomposed into a set of scale images and each scale image consists of square texels of the same size. The texels in a scale image may have different pixel values. The degree of presence of a texel in a textural image can be measured by the image area occupied by the texel in terms of pixel. The histogram of texel area is shown to be a useful texture feature, and a dominant texture scale derived from the histogram provides a good reference parameter for computing gray-level co-occurrence matrix.

Effects of Apparatus Parameters on MFL Signals Using Orthogonal Experimental Design

Magnetic flux leakage (MFL) is a non-destructive testing method used to inspect ferrous materials. However, there are a variety of factors that can affect the MFL inspection tool’s ability to detect and characterize anomalies. An orthogonal experimental design (OED) method is applied to study the effects of apparatus parameters on MFL signals. Integration of OED method of analysis into a routine sample preparation technique could improve the repeatability and quantization capabilities of MFL tools. Three key apparatus parameters, namely permanent magnet (PM) height, magnetic concentrator (MC) length and backing iron (BI) length are chosen for the present study. The importance of each of these key parameters on MFL signals for different defects is determined by a series of experiments.

Illumination Invariant Texture Classification with Pattern Co-occurrence Matrix

A new illumination invariant feature extraction method for texture classification is proposed. In order to capture the local image texture, texture pattern transform (TPT) in a local neighborhood of a monochrome texture image is introduced. The TPT is robust against any monotonic transformation of the gray scale. The joint distributions of two different TPT, which can be characterized using a pattern co-occurrence matrix (PCM), can be used for texture classification. The PCM technique only requires comparison and counting operations, and thus is highly computationally efficient. The properties of PCM include translation and illuminant invariance, which is highly desirable in real-world applications. Illumination invariant texture classification experimental results show that the texture features derived from PCM achieve good discrimination.

Data Fusion for MFL Signal Characterization

The objective of data fusion is to be able to draw inferences that may not be feasible with data from a single sensor alone. In this paper, data from three sets of sensors are fused to estimate the defect profile from magnetic flux leakage (MFL) inspection data. The three sensors measure the axial, radial and tangential components of the MFL field. Data is fused at the feature level. Examples of signal features are amplitude, width, etc. A radial basis function network (RBFN) is then employed to map the fused features appropriately to obtain the geometric profile of the defect. The feasibility of the approach is evaluated using the data obtained from the MFL inspection of oil pipes. The results obtained by fusing the axial, radial and tangential components appear to be better than those obtained using the axial and radial component alone.

Numerical Simulation and Analysis on Magnetizing Exciter for Magnetic Flux Leakage

Magnetic flux leakage (MFL) is a non-destructive testing method used to inspect the pipe and magnetization of the pipe wall to saturation is essential for anomalies to be reliably and accurately detected and characterized. Axial components of magnetic flux density obtained during the MFL inspection have been simulated using three-dimensional finite element analysis and the effects of magnetizing exciter parameters on magnetic flux density are investigated. The pipe modeled in this paper has an outer diameter of 127mm (5 in.) with a wall thickness of 9 mm (0.354 in.). According to numerical simulations, an increase in the magnetic flux density of pipe wall is observed with an increase in the permanent magnet length and height. It clearly illustrates that Nd-Fe-B permanent magnet assembly with 70 mm length and 40 mm height may magnetize pipe wall to near saturation.

Affine Texture Analysis with Scale-Area Histogram

A major problem of texture analysis is that textures in the real world are often not uniform due to variations in orientation, scale, or other visual appearance. In this paper, affine texture analysis with texel scale-area histogram is presented. A textural image is decomposed into a set of scale images and each scale image consists of square texels of the same size. The scale-area histogram of texel is used as texture feature for multi-scale texture analysis and dominant texture scale analysis. Measurement of the dominant texel sizes of textural images with different rotation angles and spatial scales indicates that rotational and scaled transformations of textural image result in the motion of translation in scale-area histogram.

Automated Visual Inspection of Surface Defects on Hot-Rolled Plate

This paper is concerned with the problem of automatic inspection of hot-rolled plate surface using machine vision. An automated visual inspection (AVI) system has been developed to take images of external hot-rolled plate surfaces and the detailed characteristics of the sensor system which include the illumination and digital camera are described. An intelligent surface defect detection paradigm based on morphology is proposed to detect structural defects on plate surfaces. The proposed method has been implemented and tested on a number of hot-rolled plate surfaces. The results suggest that the method can provide an accurate identification to the defects and can be developed into a commercial visual inspection system.

Surface Inspection Using Computer Vision and Gradient Spectrum

This paper is concerned with the problem of automatic inspection of hot-rolled plate surface using computer vision. An automated visual inspection system has been developed to take images of external hot-rolled plate surfaces and an intelligent surface defect detection paradigm based on gradient spectrum technique is presented. Gradient spectrum characterizes the spatial configuration of local image texture and is robust against any monotonic transformation of the gray scale. Texture features based on gradient spectrum are extracted from ROI in hot-rolled plate surface images and integrated into a feature vector which uniquely differentiates the abnormal regions from normal surface. Classification accuracies using the gradient spectrum and gradient-based method are compared. The results indicate that gradient spectrum performs well in classifying the samples with the lowest classification error.

Effects of Magnetic Concentrator on MFL Signals Using FEA

Magnetic flux leakage (MFL) is a non-destructive testing method used to inspect ferrous materials. However, there are a variety of factors that can affect the MFL inspection tool’s ability to detect and characterize anomalies. MFL signals obtained during the inspection of pipes have been simulated using 3D finite element analysis (FEA) and the effects of magnetic concentrator on MFL signals are investigated. Measurements of the leakage flux with various defect depths or widths indicate that the axial component of MFL are improved by magnetic concentrator with the result that significant advantages could be obtained in defect detection schemes, in that the signal to noise ratio (SNR) of MFL signals can be improved by magnetic concentrator.

3D Finite Element Analysis for Magnetic Flux Leakage Testing

Magnetic flux leakage (MFL) is a non-destructive testing method used to inspect ferrous materials. However, apparatus parameters could affect the MFL inspection tool’s ability to characterize anomalies. In this paper, MFL signals obtained during the inspection of pipes have been simulated using three-dimensional finite element analysis and the effects of magnet assembly on MFL signals are investigated. According to numerical simulations, an increase in the leakage flux amplitude is observed with an increase in the permanent magnet size and the inflexion point may indicate the presence of magnetizing pipe wall to near saturation. It clearly illustrates degradation in the MFL with increasing backing iron length. The relationship between MFL apparatus parameters and MFL signals could be utilized in the MFL technique to characterize the defect.