以影像處理自動化量測顯示器畫質

Automatic Display Quality Measurement by Image Processing

陳柏升(Bo-Sheng Chen)

指導教授 ： 謝禎冏

大同大學/電機資訊學院/資訊工程學系所/碩士(2009年)

摘要

本論文提出一個以影像處理自動評量顯示器畫質各種規格參數的方法，以電腦視覺代替人眼，客觀評鑑面版品質，藉此讓消費者選購時可以有參考的依據。一般人眼容易觀察到的顯示器基本表現參數有顏色、對比、亮度、銳利度等四種。除了靜態的影像參數之外，動態影像模糊(Motion Blur)亦是影響顯示器好壞的重要因素之一，動態模糊是螢幕在播放快速畫面時，會因液晶扭轉速度不夠，造成移動物件的四周產生殘影。本論文提出利用一般的網路攝影機及DV，量測上述五種螢幕參數的方法，對於不同的參數，我們提出了不同的樣本與方法來進行測量，依不同樣本的圖像特性，以影像處理技術取得其特徵，得到電腦初步評估分數，之後再經由線性迴歸分析，推算出人眼大約估算的分數，藉此來分辨顯示器的好壞。
本論文收集10台不同廠牌的液晶螢幕，每台由10個人以及本論文所提之系統來做評估，透過迴歸分析方程式，修正本系統與人眼評比之間的差異。之後，本系統評估另外3台顯示器，將評估的分數與人眼評估做比較，發現兩者之間的誤差在1.8%以下，證明本論文所提出之方法確實可行。

關鍵字

迴歸分析 ； 動態模糊 ； 影像品質 ； 自動測量

並列摘要

This thesis develops a system for automatic display quality measurement by image processing. The goal is to replace human eyes for display quality evaluation by computer vision and get the objective quality review that could be referenced for consumer to make purchase. Color, contrast, brightness, and sharpness, are the main properties which could be easily recognized by human vision. Besides of static image parameters, motion blur is one of the most important factors to affect display quality. Due to the limited reversal speed of liquid crystal, there is the blur around the moving object when displaying fast moving objects. In this paper, common webcam and DV are deployed to measure the above mentioned five display quality parameters. Different patterns displayed on the screen are captured and analyzed by image processing techniques to get the image characteristic. A score is calculated by the measured image characteristic. Then, linear regression model is adopted to find the relation between human score and the measured display performance.
In the experiments, ten different brands LCD monitors are evaluated by persons and our automatic measurement system. Linear regression formulas are learned to find the corresponding human vision score by giving computer vision score. Finally, three new monitors are tested by our system. The results are compared with human evaluation. The errors are under 1.8% in average which demonstrated the feasibility of proposed system.

並列關鍵字

Automatic Measurement ； Image Quality ； Regression Analysis ； Motion Blur

參考文獻

[2] K. Hirai, T. Nakaguchi, N. Tsumura and Y. Miyake*,” Correlation Analysis between Motion Blur Width and Human Perception”. IDW'07, 2007

[4] Isao Kawahara.” Measurement of Moving Picture Resolution for Displays Using Scrolled Sine-Bursts”, pp.121-124 , SID 08 DIGEST.

[5] A. M. Eskicioglu and P. S. Fisher, “Image quality measures and their performance,” IEEE Trans. Communications, vol. 43, pp. 2959–2965, Dec. 1995.

[6] 洪國敦,”色域映對理論應用於顯示器色彩調校之研究”, 2006.

[7] Z. Wang, A. C. Bovik, and L. Lu, “Why is image quality assessment so difficult?.” accepted by IEEE Int. Conf. Acoust., Speech, and Signal Processing, May 2002.