期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：14
页码：3312
出版社：Journal of Theoretical and Applied
摘要：Video data size could become a problem when it is accessed or stored on mobile devices. Currently, converting video into image mosaicing is the best solution for video data storage on mobile devices. However, underwater video images have more complicated problems. They have common concerns, such as low quality image. The deeper the underwater object is, the more unclear its image become. Although lighting is relatively better in shallow waters, images obtained here often contains flicker, which happens when sunlight is reflected on the water’s surface. Therefore, in the mosaicing process, feature detection must be done first before it is registered. This study attempts to make improvements to underwater image registration process by eliminating flicker so that the feature of the image can be registered. The proposed solution is to use a method to perform flicker detection and removal based on the value (v) in the HSV color model. The method will remove the flicker by automatic selection best value. From the exsperiment result shows 0.65 as the best value (v) for reference flicker. On these values (v), the amount of features that can be detected and registered using SIFT are always higher than using other value (v). Thus, this proposed method can contribute to the image mosaicing registration process
