基于小波分解和信息熵的涡结构识别方法

Vortex structures recognition based on wavelet decomposition and entropy

利用涡结构对光波通过高速流场的光学传输特性进行建模,就必须进行涡结构的识别。文中提出一种涡结构识别方法,首先把密度场转换的折射率场等效为具有丰富纹理信息的灰度图像,进而对其进行多小波分解,计算每次分解的小波基系数的熵值,利用阈值判定低熵的为大涡的小波基系数,高熵的为小涡的小波基系数。计算机仿真结果表明,该方法能有效地进行涡结构识别,为基于涡结构的光学传输精确建模奠定了基础。

It＇s necessary to recognize the different vortices, when aero-optical distortion caused by hypersonic flow is studied using vortex structure as the turbulence model. For this purpose, a recognized method is proposed. Firstly, the turbulence density field is transformed to refractive index field, which is equal to gray scale images with abundant texture information. Then wavelet transform is applied to decompose these images and the wavelet based coefficient entropies are calculated. By comparing the threshold, the coefficients with low entropy are considered to be associated with large-scale vortices＇ while the one with the high entropy are related to small-scale vortices＇. The computer simulation shows that the proposed method is valid using vortex structures recognition, and it provides basis for optical propagation model based on turbulence vortex structures.