Characteristics of spectral illumination and automatic feature inspection for stem accessory

Abstract

The study investigates the characteristics of spectral illumination and automatic feature inspection for vision image of stem accessory. The angle, intensity, and spectral analyses of light sources are analyzed for image inspections. The geometry size, roundness, and image stitching of the stem accessory are recognized for feature inspections using image morphology. For spectral illumination of white light LED arrays at various shift displacements, the maximum errors of 0, 20, 30 and 40 degrees that are compared to the shift displacement of each 0 cm are 4.2, 7.8, 6.8, and 8.1%, respectively. The deviation errors of image stitching for stem accessory in x and y coordinates are 2 pixels. The SIFT and RANSAC enable to transform the stem image into local feature coordinates that are invariant to the illumination change. A white balance is typically achieved by using correction filters of GWA algorithm over the lights or on the camera lens. Therefore, the image inspections for object recognition are depended on various spectral illuminations.