Research On Massive Remote Sensing Data Retrieval Technology Based On Spatial Grid Filtering

With the rapid development of high-resolution earth observation technology,remote sensing(RS)has been widely used in China's agricultural and forestry resource investigation,environmental monitoring of natural disasters,geology and mineral resources exploration,water resources protection,etc.The data amount of RS is also increasing day by day,which has shown a trend of big data and mass information.In order to meet the increasing demand of massive RS data,how to find the required data quickly and accurately is becoming one of the problems in research of RS data organization and management.Although there has been some achievements in research of RS data organization,management and retrieval all over the world,problems need to be solved still exist.It is a common requirement for RS technology to produce a whole map quickly in use of the monitored RS data,and a full coverage search is exactly an important part of the map making.Most of the current researches are concerned about the regional coverage with RS image,but rarely with respect to the single time phase full coverage search.In order to solve this problem,this thesis presents a data retrieval method for huge number of RS data based on spatial grid filtering.This method uses the cloud-free RS image retrieval technology based on spatial secondary filtering and the cloud-free RS image retrieval technology based on spatial grid filtering to solve the problem of full coverage retrieval.By comparing the experimental results with analysis,it can verified the feasibility of the method,as well as features of automatic,efficient and accurate in screen of the required RS data.The main contents of this paper are as follows:1)Studied the concept and segmentation standard of the Five Layers Fifteen Levels(FLFL)grid segmentation model,and its grid hierarchy selection strategy.2)Made research of the method to analyze the fast view of each RS image according to the image pixel value,and how to select the cloud threshold corresponding to each image dynamically.3)Based on the FLFL model,studied the data retrieval technology of cloudless RS image based on spatial grid,and proposed a retrieval model for huge number of RS data based on spatial grid filtering,the model can screen out a data-selected scheme with single-phase full coverage from RS images.4)Concerning the area retrieval issue of single phase full coverage,proposed a retrieval model based on spatial secondary filtering.The model can retrieve out the single phase full coverage RS data from the massive RS images quickly,automatically and accurately.5)Through the realization and verification of the prototype of the two retrieval models,the experimental results are obtained,and the results are analyzed by comparison,which shows the feasibility of the technology.In this thesis,we proposed a data retrieval method for massive RS data based on spatial grid filtering,which can realize the single time full coverage retrieval for RS data rapidly and automatically,and provide a new retrieval idea and method for searching and application of mass RS data.It can provide some theoretical basics for the RS data retrieval theory in future research and practice.