Research On The Method Of Improving Synthetic Infrared Image Fidelity And Quality Evaluation

The wide application of infrared technology in military and civilian fields makes research on infrared simulation technology becomes important.By use of infrared simulation,the infrared image can be acquired at different time and environment with low cost,which can effectively shorten the time and cost of development.The synthetic infrared image can be applied to the algorithm verification,combat training,system demonstration and other aspects.Compared with the real infrared image acquisition equipment,the simulation system is more interactive and flexible.As an alternative to the real image,the fidelity of synthetic image determines its application,so it is important to improve the fidelity of synthetic image.Based on the analysis of the factors that affect the quality of infrared simulation,the corresponding research work is carried out from three aspects: the correctness of the simulation model,the precision of the model and the sensor's effect.In the simulation model,the existing temperature model is improved,the common surface types of ground objects are divided into different types,and the corresponding surface temperature model is used to make them conform to the characteristics of material.In modeling,to solve the problem of insufficient detailed information,the visible light texture modulation method is proposed to enhance the texture details;using surface normal vector perturbation method to improve the details of rough surface;three dimensional finite element method is used to calculate the temperature field of the object.In the aspect of sensor's effect,a gain control quantization threshold adjusting method is proposed to automatically adjusting the threshold of radiant energy,which can improve image details.For different imaging effect,the corresponding model is used.The commonly used image quality evaluation method are summarized,and several methods are chosen to evaluate their applicability and reliability in assessment of infrared image quality.The simulation results are evaluated by CWSSIM method.The results were compared with commercial software Vega.The real temperature and the typical infrared image are collected and compared with the simulation results.Results show that the texture details and detector effects are the important factors affecting the simulation quality,The proposed method can improve the fidelity of the synthetic image.