Fuzzy Inspection of Fabric Defects Based on Particle Swarm Optimization (PSO)
A new approach for inspection of fabric defects based on Principal Component Analysis (PCA) and Fuzzy C-Mean Clustering (FCM) Based on Particle Swarm Optimization (PSO) is proposed. First, the PCA is used to reduce the dimension of the original image and
- PDF / 778,091 Bytes
- 7 Pages / 430 x 660 pts Page_size
- 55 Downloads / 180 Views
Abstract. A new approach for inspection of fabric defects based on Principal Component Analysis (PCA) and Fuzzy C-Mean Clustering (FCM) Based on Particle Swarm Optimization (PSO) is proposed. First, the PCA is used to reduce the dimension of the original image and computation complexity. The dimension-reduced image features, which can best describe the original image without unnecessary data, are recognized by FCM based on PSO next. The recognition is carried out by the merits of the overall optimizing and higher convergent speed of PSO combined with FCM algorithm, which makes the algorithm have a strong overall searching capacity and avoids the local minimum problems of FCM. At the same time, it reduce the degree of sensitivity of FCM that depends on the initialization values. The results show that the method is more effective than the traditional one with BP neural networks based on wavelet[1,2]. Keywords: Fabric Defects Inspection; Principal Component Analysis (PCA); Particle Swarm Optimization (PSO); Fuzzy C-Mean Clustering (FCM).
1
Introduction
Fabric defects inspection is a technology of fabric quality inspection, which use image sensor to get the defect fabric images and consider them as a template to recognize the defect fabric. Though the production efficiency of actual textile increases tremendously, fabric defects are still inspected with manual work. This method has many disadvantages, such as low speed, high mistakes and leaks, wrong result due to the inspector’s subjective consciousness, and so on. On account of this, the technology of fabric defects inspection is a urgent need for improvement in weave industry. Since the 70s of the 20th century, many domestic and foreign experts have done some researches about fabric defects inspection and educed some effective methods. However, it is on the stating stage, and most of them are reported as algorithm, patent and sample machine. Until now, adult fabric defects autoinspection system begin to enter into market[3]. With the wider and wider use of image sensor, adoption of image information for fabric defects inspection becomes one of the head methods. The image method has two species: one is G. Wang et al. (Eds.): RSKT 2008, LNAI 5009, pp. 700–706, 2008. c Springer-Verlag Berlin Heidelberg 2008
Fuzzy Inspection of Fabric Defects Based on PSO
701
processed in time domain directly, like eigenvalue picking-up from spatial texture features. Another is in frequency domain just as Fourier transform, Gabor filter or wavelet transform[4]. In our paper, a new approach for fabric defects inspection based on Principal Component Analysis (PCA) and Fuzzy C-Mean Clustering (FCM) Based on Particle Swarm Optimization (PSO) is presented.
2
Principal Component Analysis
Principal Component Analysis is a common, simple and effective linear analysis method, which is proposed by Turkm and Pentland in the 90s of the 20th century[5]. It transforms the images by Karhunen-Loeve method based on their statistical speciality, then considers feature matrix that is composed of eig
Data Loading...
