China's LCD panel production capacity has ranked the second in the world, thus China has become an industry leader in flat panel displays. With the increasing size and resolution of displays and the rapid development of new display technologies (OLED), technical requirements for the detection and correction of various flat panel pixel level defects are also constantly increasing. How to improve the high-precision detection rate of the uneven signal (Mura) is is a core issue that are of most significance in screen inspection field currently.

Mura detection is different from general pixel dead point and scratch defect detection. It is the light and dark color difference introduced in the LCD/OLED manufacturing proccess, with no obvious detection means, therefore simply raising the camera resolution cannot achieve the detection requirements of lower contrast Mura. . The Indigo Imaging team innovatively introduced a signal quantization recognition algorithm and successfully developed a high-resolution camera with ultra-high defect detection rate – the INDIGO 4500, which can not only realize a micro defect detection of 5 microns, but also can reach an ultra-low level of 1%, far beyond the industry standards.
 

As shown in the figure, the INDIGO 4500 adopts a new technology mode. The oscillation amplitude of INDIGO 4500 signal curve (Figure B) has reduced largely, the same as its quantization error, as a result the Mura signal recognition accuracy has been significantly improved compared with the Mura signal curve (Figure A) detected by the mainstream machine vision cameras on the market; On the other hand, based on the temporal and spatial correlation of the noise, the camera accomplishes a multiple iterations of pre-processing, under the premise of not affecting the real signal, to separate the signal to be measured and the jitter of the background better, as shown in Figure C: Compared with the un-preprocessed image, the signal-to-noise ratio of the INDIGO 4500 is improved by more than 2 times after pre-processing, and the sensitivity of the Mura signal detection can be improved by 78% on the original basis, which realizes the precise detection of micro-signals.