Research On Scale Effort And Optimal Scale For Remote Sensing Image

With the rapid development and wide application of satellite remote sensing technology,the resolution of remote sensing image has been put forward higher requirements,which also results to the exponential growth of image data.However,by the existing wireless communication technology,the satellite image data can not be spread quickly due to the limited transmission capacity.The large amount of image data also occupy a large amount of storage space,processing time and transmission bandwidth,all of these reasons have seriously restricted the further development of remote sensing technology.In fact,not all information in remote sensing image is effective,some may be redundant.Therefore,a new research area has been put forward,which is the scale of remote sensing image.Scale refers to the distance that pixels can represent.Different research purposes also need different scales of remote sensing images.This paper will select the best spatial scale for camera sensors on remote sensing satellites.The best spatial scale sensing images can distinguish different object types from the target scene,display the spatial distribution rules of the objects,and the images have a larger width,the details of the images are not too complicated.The scale research of remote sensing image mainly includes three aspects:scale conversion,scale effect and the choice of optimal scale.This paper will take the remote sensing images whose scales ars already known as research object,and use different methods to do conversion,then study the scale effect of these processes.At last,this paper will choose the optimal scale for various scenes.The research details are as follows.Firstly,this paper will introduce the necessity of scale conversion and several commonly used scale conversion methods,also present the advantages and disadvantages for different conversion methods.Then,three scale conversion methods will be used to generate the research samples.Secondly,we will study the scale effect caused by these three scale conversion methods using the specific scale effect evaluation index:mean of gray,standard deviation of gray and spatial correlation coefficient.It will be found that the bilinear interpolation conversion method is the best scale conversion method.And this result will be verified by actual ground scene experiment.This paper will use this method to generate a series of images for further study.Finally,the optimal scale will be selected for each scene.In this study process,the dynamic range and the image entropy are used as the reference factors.And the spatial correlation between pixels will be investigated by average of local variance and the variant function method.