Applying the support vector machine with optimal parameter design into an automatic inspection system for classifying micro-defects on surfaces of light-emitting diode chips

A Review and Analysis of Automatic Optical Inspection and Quality Monitoring Methods in Electronics Industry

“…These defects were analysed geometrically by evaluating area and perimetric measures of the segmented areas of interest acquired from images. Kuo et al in [46] investigated LED chip defects in two areas of the chip; electrode area and light area, which directly affect the luminous efficiency (shown in Figure 7(a)). The investigated defects in electrode area in this study are contamination, non-probe and scrape as shown in Figure 7(b).…”

“…[49] Luminance and forward voltage inspection LED chips [44] [50] Inspection of LED die defects geometrically (in terms of die size, electrode size) and heuristically in terms of other defect types in lightemitting regions and electrode region such as empty points in light-emitting region [36] [51] Water-drop defects that causes blemishes on LED chip surface [52] Overall inspection of LED dies in a wafer in terms of the quality of light-emitted area, Nelectrode, P-electrode and probe marks [52] Overall inspection of LED dies in a wafer in terms of the quality of light-emitted area, Nelectrode, P-electrode and probe marks [20] Discontinuity and erosion defects located at the area of p-electrodes and disjunction region of LED die. [45] LED chip defects that include: fragment chips, scratch marks and remained gold on the pad area, scratch marks on the luminous zone and missing luminous zone [35] LED wafer luminance test [48] Polycrystalline and fragmentary defects [46] Electrode area defects such as contamination, scrap, and non-probe defects. Light area defects such as breakdown and color aberration defects [47] Line blemishes and scratch marks in two types of chips face [54].…”

“…In other words, if the positive class is smaller than the negative class, then the hyperplane will move toward the positive class, which will further result in numerous false negative errors [269]. Kuo et al in [46] investigated two types of defects in the light area and three types of defects in the electrode area of LED chip using several SVM-based algorithms. Since SVM is optimized for binary classification, a standalone SVM algorithm were used to classify light area defects into two defect classes: breakdown and color aberration.…”

Przegląd Wykorzystania Aoi W Procesie Kontroli Montażu Powierzchniowego

“…In [18], an optimal design of an automatic inspection system for processing light-emitting diode (LED) chips was discussed. The system is based on a support vector machine (SVM).…”

A Research of Micromachining of Materials

“…The instrument deflection from the micromilling influences precision and the processor development of this surface when compared with traditional milling. The application edges) n d I s n I f o r m y combined the border are tremendously essential whilst the processor depth becomes a comparable into this border radius [7]. Due to the fact the processor load is modest when compared with this cutting border radius, the measurement effect also plowing compels be important on surface and induce production.…”