Visual Inspection of Machined Metallic High-Precision Surfaces
- PDF / 1,571,051 Bytes
- 12 Pages / 612 x 792 pts (letter) Page_size
- 100 Downloads / 179 Views
Visual Inspection of Machined Metallic High-Precision Surfaces Franz Pernkopf Institute of Automation, University of Leoben, Peter-Tunner-Strasse 27, A 8700 Leoben, Austria Email: [email protected]
Paul O’Leary Institute of Automation, University of Leoben, Peter-Tunner-Strasse 27, A 8700 Leoben, Austria Email: [email protected] Received 31 July 2001 and in revised form 12 January 2002 This paper presents a surface inspection prototype of an automatic system for precision ground metallic surfaces, in this case bearing rolls. The surface reflectance properties are modeled and verified with optical experiments. The aim being to determine the optical arrangement for illumination and observation, where the contrast between errors and intact surface is maximized. A new adaptive threshold selection algorithm for segmentation is presented. Additionally, is included an evaluation of a large number of published sequential search algorithms for selection of the best subset of features for the classification with a comparison of their computational requirements. Finally, the results of classification for 540 flaw images are presented. Keywords and phrases: visual inspection, surface reflection, flaw detection, segmentation, statistical classification, feature selection.
1.
INTRODUCTION
In industry, there is an increasing demand for automatic surface inspection systems for quality control of final products. Thereby, image processing techniques play a crucial role in this growing field. The customer demands are well founded on the high costs of poor quality, with the resulting cost for correction. The human assessment of the product quality in repetitive tasks, for example, inspection of textiles, achieves a maximum reliability of only 80% [1]. Thus, replacing the human visual inspection with a fully automatic system will normally increase the accuracy of flaw detection and reduce costs. A survey of automated visual inspection can be found in [2]. Many papers of surface inspection on steel products have recently been published [3, 4], most of which concentrate on flat objects. In this paper, an automatic visual surface inspection system for bearing rolls, that is, precision surfaces, is presented. A coarse concept of the automated solution for real-time visual inspection of bearing rolls is shown in Figure 1. It consists of five main parts: (1) an optical arrangement for the acquisition of the surface image data, including the mechanical hardware, camera, lighting, and frame-grabber (see Section 3), (2) flaw detection (see Section 4), (3) image segmentation (see Section 5.1),
(4) feature extraction (see Section 5.2), (5) feature selection (see Section 5.3), and (6) classification (see Section 5.4). The problem at hand is presented in Section 2, whereby, the demands on the inspection system and the main characteristics of the flaws which must be detected are described. The design of the optical arrangement (see Section 3) is based on a modeling of the surface reflectance properties [5, 6, 7]. Exploratory
Data Loading...
