5441337092

Figurę 7. Misalarm rates comparison.

IC

QFP

IC

EGA

The histogram of the images of the solder pastę and PCB base is shownin Rg. 6, Rg. 6 (a) is the histogram of the image of the PCB base. Rg. 6 (b) i s the histogram of the image of the solder pastę before image enhancement Rg. 6(c) is the hi stogram of image c£ solder pastę after image enhancement. It can be seen that the histogram value of solder pastę image was increased after image enhancement, and the differenoe of the mean grayscale values between the two images yaries from 4.6 (109.8-105.2) tol3.7 (118.9-105.2). The increase value was related to the weight factor h.

(a) Histogram of the image of the PCB base

~[Mean of gaysoale: 109.

Rg. 7 shows the misalarm rates cf solder pastę depositing defects for four kinds of electranic component Rem this figurę, one finds that the misalarm rates reduced by using position compensation and teocture-based image enhancement method.

TABLEI Mb alarm Rates

Defeot lype of solder pastę JJg    ^

Displacement    3    A

Defidenoy    1    6

Bridge,    0    0

Oyerflow    0    1

Table I shows the misalarm rates of fiye defects of solder pastę depositing for four electronic components, two kinds chip and two kinds of [Cs. One can finds that the misalarm ratę of tire displacement defect is the highest and is related to threshold yalue set by user. Whilst, the defect misalarm ratę of the smali and round pads is hi gher.

The results of pseudo-3D inspeefion are shown in Rg. 8. Using pseudo-3D inspeefion algorithm measure the chip 1005 and chip 0402 fifteen fimes. The results show firat the fluctuation rangę of the vdume ratio described in equation (2) is from 50 to 110. Therefare, it can be ccmcluded that the 3D inspeefion method i s stable.

(b) Histogram of the image of the solder pastę before image enhan cement

(c) histOjjain ofimageofsolder pastę after image enhancement

Figurę 6. Histogram of fhe image of the solder pasie

limiled lo: Brigham Young I

Rgure 8. Pseudo-3D InspeeSon Resulis

140