Measurement of the visual perceptibility of sink marks on injection molding parts by a new fast processing model

Defect or Design? Leveraging the Angle of Opportunity for Detecting Scratches on Brushed Aluminium Surfaces

“…Designers of a semi-automated system in this case would need a resolution in rotation of 1°or higher when capturing data in order to use AoO as a defect descriptor. Especially highly reflective materials with brush strokes requires units to be inspected from various view angles but previous approaches regularly captured data from only a single viewpoint [14], [15], [17], [19], [26] and generally lacked multi-view assessment of defects taking advantage of the spatio-temporal domain [20]. Capturing single viewpoint means that defects only visible in certain view angles will not guarantee being captured and the number of false positives for a single frame can be numerous (see Figure 7).…”

“…Nondestructive methods, as machine vision, can be used to inspect surfaces without permanently altering the inspected object and are exceedingly requested in industry [11]. Previous work involving machine vision for visual inspection focused on different materials including leather [14], ceramic [15], stone [16], plastic [17], and metals [18], [19]. Searching for defects on diffuse materials as leather and stone is easier than detecting defects on reflective surfaces as metal and plastic due to the specular reflections causing overexposure and appearance variance from certain viewpoints.…”

“…Due to the vast amount of different defect types previous approaches reduce the defect space to a limited amount defect types e.g. welding, exfoliation, or sink marks [17], [18] and perform binary classification (defect versus non-defect) [31] and/or anomaly detection [36]. In other words, previous work limit the amount of classified defect types or apply anomaly detection looking for anything in-homogeneous on the surface.…”

Aesthetic defect characterization of a polymeric polarizer via structured light illumination

“…The defect model and imaging mechanism built by Seulin in 2002 were sequentially adopted by other researchers [13][14][15][16][17]. They modeled the surface defects of specular objects as bulges and dents.…”

“…Macher and Gruber used phase measuring deflectometry with structured lighting to detect visually perceptible sink marks-a type of surface defects-on high gloss injection molded parts. They then modeled the surface defects for fast measurements of their visual perceptibility [16][17].…”