| Summary: | 碩士 === 國立雲林科技大學 === 電子與光電工程研究所碩士班 === 102 === As technology advances, it is an urgent matter how to improve the dynamic range of the monitor screen. Because the dynamic range of surveillance cameras is much lower than the real-world, there will miss some of important information. Current surveillance systems often rely on automatic exposure control to obtain the image, so that they have image darker and unclear at night. In order to make the image brighter and clearer at night, surveillance cameras are necessary to proceed long exposure time during shooting, but this will lead to moving objects blurring. The existing image exposure fusion technology can‟t solve above these problems. Therefore, we propose a novel image exposure fusion algorithm that uses block-based edge strength to find the optimal blocks from two differently exposed images. Next, a high dynamic range image is fused from these blocks. The experimental results show that the proposed algorithm can save 12.5% memory capacity, increase 13.74% computing speed, improve about 10.35% image quality of moving objects in dynamic scenes and enhance about 5.37% image quality of static scenes when comparing with the existing methods. In short, our approach possesses lower computing complexity, memory requirement and high image quality such that it is suitable for implementation in the general embedded platform.
|
|---|
