摘要
文章针对嵌入式平台实时性要求高的特点,提出一种基于运动自适应的快速去隔行算法。运动检测部分采用改进的4场检测方法,检测结果更加平滑。对于检测为静止的像素采用场复制插值法,保留图像细节,对于运动像素则采用分级边沿平均插值法,克服了传统的边沿平均插值法固有的缺点。
To satisfy the real-time requirement of the embedded system,an adaptive motion based fast deinterlacing algorithm is proposed in this paper.The improved four-field motion detection method is adopted to get more smoothing detection results.In the detected static area,field duplication is used to keep more image details,while in the detected motional area,ranked edge-based line average(RELA) method is used to overcome the intrinsic defect of traditional edge-based line average(ELA) method.
出处
《合肥工业大学学报(自然科学版)》
CAS
CSCD
北大核心
2010年第12期1802-1805,共4页
Journal of Hefei University of Technology:Natural Science
基金
国家自然科学基金资助项目(60474035)
高等学校博士学科点专项科研基金资助项目(20060359004)
安徽省自然科学基金资助项目(070412035)
关键词
运动检测
去隔行
边沿线平均
分级
motion detection
deinterlacing
edge-based line average(ELA)
rank