LMK DeMURA - High-Precision Display Nonuniformity Correction & Pixel Registration

In this contribution, we present typical practical implications of high magnification lenses required for camera-based microdisplay measurements and analyze their impact on existing measurement methods for DeMURA, resolution and contrast. Furthermore, we show hardware and software-based methods to improve or handle the shortcomings of these high-magnification lenses.

This paper provides an overview of image stitching and its general advantages and challenges. Further, we introduce a novel stitching concept based on our advanced pixel registration (APR) procedure. It allows easy and comparable flexible stitching setups for DeMURA and uniformity measurements in laboratory and production environments.

Modern single-pixel emitter displays such as OLED, MicroLEDs and LEDs suffer from production-related non-uniformity. The luminance and chromaticity can vary locally from pixel to pixel, resulting in a high-frequency non-uniformity and globally leading to a low-frequency non-uniformity. In order to correct these effects, luminance data of individual subpixels need to be measured. However, this is a very challenging and time-consuming task, especially for modern high-resolution displays.
To ensure a correct pixel registration (assigning the luminance to the correct pixel) in state of the art methods, display pixels are partially switched off [Patent US9135851B2]. However, this reduces cycle time and changes the average pixel level, which can affect the results.
We present a method to overcome these issues, called Advanced Pixel Registration (APR). It is based on a specific registration pattern applied during a teach-in process. An example pattern is provided in Figure 1 (left). After this initial registration, DeMURA measurements can be performed with only one image capture per input signal. The same is true for following displays during EOL testing, as small misalignments, which occur in production control environments as slight shifts, inclinations or rotations of the DUT (see Figure 2) can be corrected automatically.
This contribution validates the APR method using a flat and free-form curved display with methods similar to [] Feng, X. (2019), 78-2: Measurement and Evaluation of Subpixel Brightness for Demura. SID Symposium Digest of Technical Papers, 50: 1122-1125.]. The results show that the APR method can significantly improve the efficiency of DeMURA processes required for high-quality LED, OLED and MicroLED displays, regardless of their shape.

Evaluation of single emitter-based display technologies like OLED and μLED at modern display resolutions requires high-resolution measurements. The typically used oversampling often negatively affect cycle times and ILMD complexity. In this contribution, we present, explain, and validate an alternative to performing high-resolution measurements despite the Moiré phenomenon.

In this contribution, we present and validate a DeMURA procedure using a one-shot approach that does not require massive oversampling. It bases on on-site calibration of the imaging condition in the setup by a specific teach pattern. It has only to be performed for the first sample. Slight misalignments of following displays are then automatically corrected. After that, the luminance of each display can be measured with a single measurement. The proposed method has the potential to facilitate high-precision calibration of pixel-level luminance under much more relaxed sampling conditions and higher speed compared to current methods.