The Research On Models And Methods For The Visual Effect Enhancement Of Satellite Remote Sensing Imagery

With the development and innovation of satellite remote sensing technology and theory,as the carrier of 3D world space and ground object information,remote sensing imagery has been raised to the problem of "vision".The application of satellite remote sensing imagery is from 2D to 3D map,how to gain accurate geometric imaging and the embodiment of the real image visual effect,the construction of physical properties such as image geometry construction,color expression and brightness distribution,surface texture construction,is the key.How to obtain the visual effects of human subjective consistency through the geometric construction of remote sensing imagery,color synthesis and atmospheric cloud correction is closely related.The overall cognition of the visual system of human body is reflected in brightness distribution and dominant hue.The dominant hue refers to the overall color representation formed by the influence of environment and the color of light source.Because the imaging system is affected by atmosphere,light source,sensor and geometric orientation,it is easy to result in uneven distribution of brightness space of resulted imaging.The image statistical distribution is not fully matched with the dynamic range of imaging equipment,which results in the difference of the overall image color or the difference of the main color of adjacent images.The remote sensing imagery is the human visual representation of the ground features,presenting the visual information with the sensory superiority and visual immediacy.At present,many researches from the perspective of traditional image processing model are discussed,these research results can effectively solve the problem such as imaging accuracy,improve image quality,but most of the method based on the imaging model,image color synthesis and atmospheric correction without considering sensors and the difference of the perceived world.The purpose of this study is to reconstruct the visual effects and improve the texture structure of imagery through the processing models.This paper focus on the main factors influencing the visual effects of satellite remote sensing imagery,reviews the current research status at home and abroad,the development trend and existing problems,and pays more attention to the research and discussion on the construction and solving methods of the visual effects enhancement.1)Propose the construction of image imaging model embedded in PCI-RPC and ArcGIS-Spline.Based on the features of remote sensing imagery,analyze in detail the reason of geometric deformation and the theory of existing image correction model.For the mountains with high topography and terrain fluctuation,it is difficult to guarantee the control point precision for the existing correction model,propose a new image correction model merging PCI-RPC and ArcGIS-Spline,image processing software PCI combined with rational function coefficients(RPC)get the PCI-RPC model,software ArcGIS combined with Spline function get the AcrGIS-Spline model.Taking the experimental data CBERDS-02 C as an example,correct the data by the PCI-RPC model,acrgis-spline model and the imaging model based on PCI-RPC and ArcGIS-Spline,obtain three groups of corrected results.Using the precision evaluation on the control points of imagery results for the accuracy comparison,analysis and verification,it is concluded that the new correction method of remote sensing imagery based on PCI-RPC and ArcGIS-Spline model can effectively improve the image precision for complex terrains and mountains,improve the quality of remote sensing imagery,enhance the visual effects of remote sensing imagery.The same applies to pingyuan,hilly areas,with a wide-ranging nature.2)Propose the construction model of color synthesis based on the near-infrared transformation of RGB-NIRAccording the problems of different remote sensing image synthesis that the spatial texture information is not colorful,the object boundary is not clear and the adjacent image color is different,propose the construction of image color synthesis algorithm and model based on optimal band combination for near-infrared(NIR),referred to as RGB-NIR transformation model.Analyze in detail the mathematical basis of building model,function relationship,color synthesis method.From four groups(8scene)different experimental data sources,complete the color synthesis experiment based the construction model by the image preprocessing of format conversion,projection transformation,orthographic correction,image registration,etc..Experimental results verified the validity of color synthesis model based on RGB-NIR,and the texture structure of result images have more full color and nature gentle than traditional model.3)Propose the construction of cloud-corrected model CLAE based on the low atmospheric effects.According to the theoretical knowledge and mathematical foundation based on the imaging mechanism of atmospheric clouds,imaging model and image degradation model,point out that it is difficult to express and describe the point spread function(PSF)of process and mechanism complexity on the image degradation and fuzzy by a unified model,propose the construction idea of cloud model based on correcting the low atmospheric effects(CLAE).Select two sensing images with different data sources and cloud cover as the experimental data,process these data using the good cloud removal models of homomorphic filtering method and wavelet transform,and proposed cloud correction model based on CLAE.By the subjective visual analysis of image results and objective quantitative evaluation of precision parameters,the evaluation results show that the construction model of cloud correction based on CLAE can well remove the thin cloud and mist effect,improve visual effect,strengthen the ability of feature interpretation and information recognition.Through the research on models and methods for the visual effect enhancement of satellite remote sensing imagery,the experimental results show that,three kinds of model and method this paper puts forward can effectively improve the image quality,enhance the reading ability of ground objects.