Saliency Detection Based Pyramid Optimization of Large Scale Satellite Image

Large-scale satellite images play an important role in disaster monitoring, ecological protection and other fields. However, due to its large size, it leads to slow browser loading and need more storage spaces. To solve the problem of spatiotemporal processing of satellite images, this paper proposes saliency detection based spatiotemporal pyramid optimization algorithm for network transmission, browser loading and visualization of large-scale satellite images. Saliency detection was firstly used to generate graphical result to analyze the area that people are easy to pay attention to in a remote sensing image, namely the significant area. Then, slice segmentation is carried out on the remote sensing image to separate the significant area and non-significant area, and the significant area is processed separately. In the experiment, cultivated land protection monitoring system based on WebGIS was used to load satellite images. Through comparison in the cultivated land protection monitoring system project, the loading speed and efficiency of the satellite image processed by the method in this paper are improved by about 14.57% compared with the satellite image without any processing. Compared with satellite images that construct a complete image pyramid, there is little difference in loading speed, but the storage efficiency of the proposed method is improved by about 11.98%. Experiments show that the satellite images processed by the method in this paper meet people's needs for image loading, and can be applied in actual project development to save computing resources and storage resources.