| Summary: | 碩士 === 大同大學 === 資訊工程學系(所) === 97 === 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.
|
|---|
