Application Of Remote Sensing Image Registration Technology In High Altitude Area

With the introduction of big data information such as computer network technology,satellite sensor technology,spaceflight technology and new generation technology,cloud computing,Internet+,Internet of things,the application of remote sensing images has gradually become intelligent and popular.The construction and production of the country is closely related to our daily travel,life and so on.However,in remote sensing images,panchromatic images have higher spatial resolution and no color spectral information;multi-spectral images have relatively lower spatial resolution,but contain rich color spectral information.For the different characteristics of both,how to adopt scientific and effective processing means to collect the superior information of both,improve the accuracy of image registration,better complete the fusion of images,more conducive to the identification of information,feature classification,application update and so on,it is one of the key research directions for the realization of intelligent and popular application of remote sensing images at present.Because of the difference of image texture,especially in high altitude area with single texture feature and less ground target information,when building image scale space,the same feature is often identified as different feature points,but different feature points are identified as the same feature the different feature points are identified as the same feature points,and together with the influence of mixed pixels and other factors,in the process of feature point search and identification,problems such as uneven feature point extraction and inconsistency often occur.For the difficulties of image registration fusion technology above,this article first describes the research background and general research status of the field of remote sensing image registration.After a basic introduction to the concepts related to image registration,the basic methods and classification of image registration technology are summarized.Focus on research and analysis of the fusion registration of the remote sensing image of GF-2 in the weak texture area of the high-altitude areas(more than 3000 meters)in our country.Because of the GF2 is a sub-meter class high-resolution remote sensing satellite.Using the pushsweep type imaging principle,the image obtained is a linear array striped image perpendicular to the orbit direction of the satellite,and the linear system of the image is higher.These characteristics make the images have strong robustness of scale-space invariance,which is beneficial to the application of scale invariant feature algorithm.But the high altitude area of our country is sparsely populated,the feature point matching is difficult,the spectral information is relatively single,the boundary shape characteristic of the ground target is not very significant,and the mismatch phenomenon is easy to appear when the feature point matching.Based on the texture characteristics of this type of area,four different texture data were selected during the experiment,and the image registration accuracy of this type of area was improved by optimizing the ground control point scheme.The experimental results show that this method effectively reduces the mismatch of feature points,and the obtained homonym points are uniformly distributed and the accuracy is higher,and the high precision registration of remote sensing images with different textures can be achieved very well.The geometric correction of remote sensing images is guaranteed to be carried out quickly and has certain practical application value.