Technology And Software Design Of GF-4 Satellite Image Quality Improvement

The GF-4 satellite is the first geostationary orbit satellite in high-resolution Earth observation system of China.It was developed by Beijing Research Institute of Mechanical and Electrical Technology,which is mainly used in fields such as disaster reduction,meteorology,forestry,and environment.In addition,It has great application in national security,environmental monitoring,weather forecast.However,high orbits,complex working environments and variation in ambient temperature cause serious image degradation.what's worse,imaging optical systems and electronic systems are susceptible to cosmic ray and atmospheric influences.Although the GF-4 satellite has a high time resolution,the spatial resolution is not enough to meet the user's needs,and the imaging quality of the weak and small targets is much lower than that of low-orbit satellites.GF-4 satellites will be affected by conditions such as weather and light,which will cause the imaging quality of some periods and regions.This paper focuses on the actual imaging problems and parameters of GF-4 satellite,and studies the GF-4 satellite imaging quality enhancement technology,which mainly includes the improvement of image contrast and resolution.In terms of contrast enhancement,this paper proposes an image enhancement algorithm based on minimum contrast distortion based on the characteristics that the human eye is more sensitive to local contrast.An image contrast model is established based on the weber-Fechner criterion.The optimization goal is to minimize the display contrast distortion,and this problem is converted into the optimal path of the directed acyclic graph of statistical features of the image.At the same time,the tone distortion in the mapping function is constraint effectively.Experimental results show that the proposed algorithm can effectively enhance the contrast of the GF-4 satellite image,so as to meet the application requirements.In order to improve the resolution of GF-4 satellite images,this paper proposes an improved example-based image super-resolution algorithm.Firstly,the PSF parameters of GF-4 is employed to construct an image training set,and adapts anchored neighborhood regression to improve the spatial resolution of GF-4 satellite images.The experimental results show that the proposed algorithm can effectively improve the spatial resolution of GF-4 satellite images and enhance the image details.In practical applications,since the resolution of the GF-4 image is 10k×10K,SDL rendering engine is employed to real-time display.The above-mentioned image enhancement algorithm is accelerated by multi-threading and block interactive processing is proposed to develop a friendly real-time large-scale satellite image interaction processing software based on QT.