Image Rapid Merging Based On Emergency Mapping Aerial Remote Sensing System

Emergencies of natural disasters, such as earthquakes, usually lead to the traffic paralysis as roads and bridges are destroyed, which also makes land mapping difficult in the stricken area. In order to get the geoinformation the first time, aerial photography is often applied to get the image that can cover the whole stricken area after merging. It helps to make out the emergency plan and disaster relief work so that people’s lives and fortunes can be saved as much as possible.When natural disasters happen, saving time means saving people’s lives. Therefore, it is a critical question of how ho get the image and complete the image merging in the shortest time while guaranteeing the image quality in the same time. The key part of image merging lies in the registration relation of adjacent images and the fusion of the image overlapping area after merging. People have made much research and achievement in the two areas. Some mature and practical methods have been put forward in image registration and fusion.Aerial remote sensing images bear high resolution, large number and mass data. The images merge simply and fast if directly use image POS data as it abandons the complicated process of registration. However, as it is constrained to the accuracy of POS system and influenced by windy, rainy and cloudy weather, the images after merging often miss something on the land or there will be obvious gaps between images, which make it fails to provide complete and accurate information for emergency dealing. Another method is to get the image based on image characteristics. Image got by this method has high conformity, which barely miss things on the land and almost have no merging gaps. However, due to its complication, large amount of calculation and cost of time, it fails to provide information in a short time.Considering the specialty of emergency aerial remote sensing, the thesis puts forward an image merging method based on the combination of image POS information and image characteristics. According to this method, the geographical coordinates are firstly got by using image POS data to complete ortho rectification and down sampling; then image mosaic is completed based on the geographical coordinates; auto matching is next processed in the overlapping areas based on image characteristics and corresponding points are determined; then the average deviation is added to the geographical coordinates that is to redefine after image merging; mosaic is to completed based on the new geographical coordinates; the mosaic images are fused to get the single navigation merging and an image that covers the whole hit area is achieved finally.