Objective Atmospheric turbulence is a key research topic in the field of atmospheric and environmental science. It greatly influences the development of aerospace, aircraft safety control, and laser communication. Random fluctuations in the refractive index of the atmosphere along the laser propagation path cause a series of transmission effects. Therefore, a laser beam projected onto a large area microcrystalline reflective film appears as irregular dynamic speckles in the far-field plane. In addition, there are constant directional moving shadows in the laser speckle images, which are caused by the path integral effect of transverse wind. Theoretically, two-dimensional (2D) laser shadow images can be used to detect the path transverse averaged wind velocity. However, due to the deformation of laser shadows and uneven transverse wind distribution, it is uncertain whether the moving velocity of laser shadows calculated using a cross-correlation algorithm can accurately reflect the moving velocity of the flow field and how the sampling frequency of images affects the calculation results. To address these problems, a new method for simulating transverse wind in atmospheric turbulence based on the dynamic phase screen theory is developed for quantitative simulation analysis. In addition, simulation results are substantially verified by experiments. We hope that our study will be helpful in the remote sensing detection of wind and other engineering applications.
Methods A complete pixel search algorithm based on normalized cross-correlation is used to calculate the displacement of laser shadows. First, the transverse wind along the laser propagation path is simulated by moving infinitely long and nonstationary phase screens. Images of lasers transmitted through the atmosphere are obtained for uniform and nonuniform transverse wind distributions. Then, the relationships between the velocity of laser shadows and wind speed are analyzed separately when the wind flow field is distributed differently. In addition, the appropriate sampling frame rates of different wind speeds are calculated. Next, a laser propagation experiment on the horizontal path is conducted, and real-time laser shadow images are taken. Finally, the calculated displacement of laser shadow images and transverse wind speed obtained using an ultrasonic anemometer are fitted to obtain their quantitative relationship. The path transverse wind velocity can be calculated directly from laser shadow images using this relation.
Results and Discussions Simulation results show that there is a linear relationship between the moving speed of laser shadows and path transverse wind speed when the distribution is uniform. Although the path transverse wind blowing from different directions introduces errors, this linear relationship still exists (Fig. 5). The shadow displacement caused by the transverse wind near the emission end is greater than the average wind speed; whereas, it is less than the average wind speed near the receiving end, that is, the influence of transverse wind on shadow displacement has a different path weight (Fig. 6). In addition, the sampling frequency of the image has a great influence on the calculation results of shadow displacement. (Fig. 7). Finally, experimental results show that the correlation coefficient between the measured and fitted wind speeds based on shadow displacement reaches 0.949, demonstrating that 2D wind vector can be obtained using laser shadow images in actual measurements. (Fig. 10).
Conclusions Simulation analysis shows that when the path transverse wind is uniformly distributed, the movement speed of laser shadows and transverse wind speed have an approximately 1∶1 linear relationship, which means that the moving velocity of laser shadows accurately reflects the moving velocity of the flow field. If the path transverse wind is in an inconsistent direction, some fitting errors are introduced, but the linear relationship between the shadow movement speed and average transverse wind speed is maintained. Moreover, the influence of path transverse wind at different positions on the shadow displacement is slightly different, causing the proportionality coefficient not to be 1. In addition, the minimum sampling frequency of CCD is estimated to ensure sufficient spatial correlation between continuous images and the accurate calculation of laser shadow displacement. Laser propagation experimental results demonstrate that 2D wind vector can be obtained using laser shadow images in actual measurements. We can quantitatively observe the motion of a 2D field and the evolution of an atmospheric vortex on a laser propagation path using laser shadow images, which is difficult to obtain using traditional methods, such as ultrasonic anemometers and wind lidar. In addition, the transverse wind is closely related to the effect of laser atmospheric transmission heat halo, and the change in wind speed in the vertical direction near the ground is related to the surface heat flow, which can be studied as a potential application direction.
