海面及天空区域的视频图像局部或全局极值点匮乏, 传统的二维经验模式分解 (BEMD)失效。本文分析了 BEMD的频率特性, 给出了高频辅助信号的构造方法, 将高频辅助信号加入原视频图像信号, 分解得到的内蕴模式分量(IMC)减去 IMC的辅助信号, 得到原信号的最接近辅助信号的频率分量。以此类推, 分解可得到原信号的不同频率分量。这种基于辅助信号的经验模式分解 (ASBEMD), 解决了局部或全局极值点匮乏的海上视频图像的分解问题, 并应用于海上降质图像的增强处理, 取得了与目前公认去雾效果较好的 HE算法一致的结果。

Lacking of local or global extrema in the area of the sea and sky of the image causes the failure of the traditional Bi-dimensional Empirical Mode Decomposition (BEMD). The frequency characteristic of BEMD is analyzed, and the construction method of the high-frequency assisted signal is given. The high-frequency assisted signal is added to the original video image signal, and then we obtain the closest frequency components to the assisted signal of the original signal by subtracting the decomposed intrinsic mode component (IMC) to the assisted signal of IMC. Moreover, the different frequency components of the original signal can be obtained. Solving the problem of lacking of local or global extrema of the maritime image, this assisted signal BEMD is applied to the enhancement of maritime degraded image, and the acquired result is consistent with that of the currently accepted HE’s algorithm.