Evaluation of methods for reduction of surface texture features

Abstract

The diagnosis of worn and damaged surfaces is an important issue in machine failure analysis and condition monitoring. Of many approaches used, image classification based on feature parameters has often proven to be particularly useful. However, large image databases can be computationally costly to analyse, and the datasets are susceptible to noise. Hence, it is essential to determine which feature parameters hold the most useful information, in order to improve the classification rate and computation time. This paper presents a performance evaluation of dimension reduction techniques currently used in pattern recognition. A comparison of three methods is conducted, in order to determine which is able to produce the best results over a large range of image datasets. The methods analysed are: Non-Linear Fisher, Principal Component Analysis and kernel Principal Component Analysis. These are then tested against four different classifiers to obtain the best combination. These classifiers are: Linear Discriminant Classifier, Quadratic Discriminant Classifier, k-Nearest Neighbour and Support Vector Machine Classifiers. For further analysis, two combined dimension reduction and classification methods are tested: Minimum Classification Error and reduced feature space Support Vectors. For the comparison, four datasets of images with different scales and rotations are used, i.e. Brodatz textures, artificially generated isotropic fractal images and Talysurf images of sandblasted and abraded steel surfaces. The results showed that a combination of the Non-Linear Fisher dimension reduction technique and a Linear Support Vector Machine Classifier gave the best performance overall and are the most promising for the application in automated machine condition monitoring and expert free failure analysis. Further improvement can be achieved by performing a step-wise dimension reduction by first reducing the features using the Principal Component Analysis method, then further reduction with the Non-Linear Fisher technique.

Shape and texture features in the automated classification of adhesive and abrasive wear particles

“…The wear particles were then classified using a Linear Support Vector Machine (Linear SVM) [33]. Detailed information on the feature extraction method can be found in [25] and on the dimension reduction technique in [26].…”

“…The classification method, developed previously [25,26] and briefly described in section 2.8, was applied to the eight classes of adhesive and abrasive wear particles As already explained in section 2.7, 50 particles from each particle class were first used to train the classification system. The performance of the classification system was then tested using 100 particle images from each class.…”

“…Secondary electron images of wear particles generated during the experiments were acquired using a scanning electron microscope (SEM). Classification was performed using a pattern recognition system developed and tested previously by the authors [25,26]. This system has been specially developed to minimize the effects of image magnification and particle orientation on classification rate.…”

Powder‐shape analysis and sintering behavior of high‐density polyethylene powders for rotational molding

“…The procedure used in this work to determine at each step length the estimated contour length is the same as that described by Allen et al16 The minimum step length was chosen to be 2 pixels, corresponding to 5.9 μm, the maximum equal to a half of the maximum Feret diameter. Further, at each step length, the perimeter was calculated starting from every point of the contour, and the average value was taken as the estimated length,33 according to eq. (8): where n is the position of the starting point on the contour and N is the total number of points in the contour.…”

“…(8): where n is the position of the starting point on the contour and N is the total number of points in the contour. In this way, the calculated estimated perimeter was made invariant from the choice of the starting point 33. Richardson plots for powders L, M, and H are shown in Figure 1.…”