Detection Image Simulation Method Of Infrared Cloud Scene Based On Fractal Theory

Space-based infrared optical satellite is an effective means to detect small scale moving targets such as moving aircraft and high speed targets in near space.However,the cloud scene in the space-based infrared observation scene has the characteristics of geometric structure dynamic change,scale dynamic change,radiation dynamic change,etc.,and is coupled with the space-based dynamic detection link,which will have a great impact on the detection efficiency of the system.Therefore,it is necessary to carry out research on the simulation methods of dynamic cloud scenes and cloud images under the conditions of space-based observation,so as to provide data support for the indicator demonstration of space-based infrared optical satellite and the design and optimization of target detection algorithm.Cloud is an important background source of space-based infrared detection and imaging.Infrared detection image simulation technology for cloud scene has always been a research hotspot in related fields at home and abroad.Most of the existing methods are only for static cloud scene image simulation or dynamic cloud image simulation of local small scenes,while few people study the large-scale dynamic cloud image simulation methods.However,modeling of large scale dynamic cloud imaging characteristics for real space-based detection conditions is an urgent requirement of space-based infrared earth detection simulation experiments.Aiming at the above problems,this paper studies the simulation method of large-scale dynamic cloud infrared detection image.The main research contents are divided into the following aspects:(1)Modeling of cloud three-dimensional spatial distribution.Based on cloud detection image simulation requirements,combined with water vapor cloud physics,from the space of cloud cover,the water content of distribution and particle size distribution of the three aspects of ice water,using fractal theory of multi-scale superposition algorithm three-dimensional spatial distribution of the cloud model is set up,mainly including the clouds appearance shape spatial distribution model of internal microstructure model and clouds.Spatial distribution modeling was carried out for different types of clouds,such as stratus,cirrus and cumulus,to provide computational support for subsequent cloud optical property calculation and radiative transport property modeling.(2)Modeling of cloud 3D radiation scene.Combined with the three-dimensional spatial distribution model of clouds,the particle distribution characteristics and optical properties of typical water clouds and cirrus clouds are analyzed,including the asymmetry factor,extinction efficiency factor and single scattering albedo of each particle at different scales.Based on the radiation transmission mechanism of cloud and the particle distribution characteristics and optical characteristics of cloud,the spherical harmonic discrete coordinate method is introduced to calculate the radiation transmission of cloud three-dimensional optical characteristics grid,so as to improve the computational efficiency of cloud three-dimensional radiation scene modeling.(3)Remote sensing chain based cloud radiation image simulation.From the perspective of optical remote sensing imaging whole link,considering the target and the background radiation energy in the atmospheric environment,optical system,detector photoelectric conversion system and imaging system,the transmission characteristics of electronics imaging link each link calculation model is set up,mainly including the model atmospheric transmission characteristics,focal plane imaging model,the platform and the load model,the signal model and so on.Combined with the cloud three-dimensional radiation scene model and the transfer function degradation model through the main link of the imaging link,the cloud radiation image simulation experiment.(4)Dynamic cloud scene image simulation based on time-series smoothing multiscale superposition method.According to the two motion modes of the dynamic cloud in the actual scene: Cloud the overall structural change in the position and shape of the clouds,through the analysis of the multi-scale superposition of traditional algorithm of dynamic cloud image simulation method,in view of the traditional methods in large scale dynamic cloud image simulation applications of low computational efficiency and large memory problems,puts forward a multi-scale superposition in temporal smoothing method of dynamic cloud image simulation method,Based on the simulation of the whole cloud position movement based on the multi-scale superposition algorithm,the shape and structure change of dynamic cloud is realized by the method of interframe interpolation,which improves the computational efficiency and realizes the simulation of large scale dynamic cloud image.