|
|
|
|
|
|
|
|
|
Low-contrast surface inspection of mura defects in liquid crystal displays
using optical flow-based motion image analysis Student：Hsin-Yang Tsai Advisor: Dr. Du-Ming Tsai Department of Industrial Engineering and Management ABSTRACT This research proposes a machine vision scheme for mura defect detection in TFT-LCD manufacturing. Mura is a Japanese word for blemish, which typically shows brightness imperfections from its surroundings in the surface. Since mura appears as a low-contrast region without clear edges in the surface, human inspectors need to continuously observe the hardly visible defect from different viewing angles. The traditional automatic visual inspection algorithms detect mura defects from individual still images. They neglect that a mura defect may not be visibly sensed in the image from a still system. In
this study, the TFT-LCD panel is assumed to move along a track, where
different light sources illuminate from different angles to the inspection
panel. While the TFT-LCD panel passes through a fixed camera, the light
reflection from different angles can effectively enhance the mura defect in the low-contrast motion images. This
research therefore proposes a motion detection scheme based on optical flow
techniques to identify mura defects in motion
images. Since the TFT-LCD moves along a single direction, both
two-dimensional (2D) and one-dimensional (1D) optical flow motion detection
methods are developed. Three discriminative features based on the flow
magnitude, mean flow magnitude and flow density in the optical flow field are
presented to extract the defective regions in each image of the motion
sequence. Both real glass substrates and synthetic panels are used to
evaluate the efficacy of the proposed inspection schemes. Experimental
results have shown that the proposed 1D optical flow method works as well as
the 2D optical flow method to detect very low-contrast mura
defects of small size, and achieves a high processing rate of 20 frames per
second for images of size 200 Keywords: Machine vision; Defect
detection; Motion images; Optical flow; Mura; TFT-LCD inspection |
|
|
|
|
《Demonstration》
|
|
|
|
|
Original images (Spot mura) |
Enhanced images |
Detection results |
|
|
|
|
|
Original images (Line mura) |
Enhanced images |
Detection results |
