摘要
图像的高频率经过预滤波和采样不能被消除,由此功率谱的振荡可能会导致混叠现象,提出了两种标准蓝噪声模式的采样方案:阶跃蓝噪声和单峰蓝噪声,两种蓝噪声在图像平面内的采样任务表现良好,在低频率段至少能与其它采样模式一样,而在高频率段不太可能导致混叠,因为它们被映射到更有效的白噪声。实验结果表明,在低采样率时采样模式能有效地防止混叠结构,在高采样率时同样表现良好。
Image of high frequency through preliminary filter sampling cannot be eliminated, and the power spectrum of the oscillation may cause the aliasing phenomenon, put forward two standard blue noise model of the sampling plan: step blue and unimodal blue noise, two blue noise in the image plane of the sample task performance is good, in the low frequency can at least with other sampling mode, and in the high frequency is unlikely to lead to aliasing, because they are mapped to the white noise more efficiently. The experimental results show that the sampling pattern can effectively prevent the mixing structure in low sampling rate, and also perform well in high sampling rate.
出处
《科技创业月刊》
2017年第12期130-131,共2页
Journal of Entrepreneurship in Science & Technology
关键词
视频流分发网络
可伸缩视频编码
多路径
video streaming distribution network
scalable video encoding
multipath