An Efficient Processing Method For Remote Sensing Image Change Detection Based On Cloud Computing

Remote sensing image change detection can meet the needs of researchers to obtain the latest information on the surface and it has far-reaching and important meanings in urban development,agricultural research,and disaster prevention.In recent years,the amount of multi-temporal remote sensing data has grown rapidly,and the traditional single-machine processing method has been unable to meet the change detection requirement of mass remote sensing images.Based on the research of OpenStack,virtualization and HDFS,this paper proposes a remote sensing big data efficient processing cloud platform based on OpenStack.In order to improve the efficiency of mass remote sensing image change detection,this paper analyzes a new change-detection method in high-resolution remote sensing images based on a conditional random field model(FCCRF-CF),and proposes a distributed and parallel method for remote sensing image change detection based on cloud computing.At the same time,in order to further improve the resource utilization of the platform,this paper proposes a multi-objective optimal scheduling algorithm based on cuckoo search.The experimental results show that the distributed parallel algorithm and scheduling algorithm proposed in this paper can realize the efficient change detection of massive remote sensing images on the cloud platform without affecting the accuracy.The main contents of this paper include:1?Proposes and designs a remote sensing big data efficient processing cloud platform based on OpenStack.This paper designs a Remote sensing image repository based on the HDFS and My SQL;componentizes the remote sensing big data processing algorithm,and combines the resource scheduling algorithm to design a business flow identification and analysis mode based on a standardized interface;based on the OpenStack API and Linux Shell scripting language,designs a cloud platform computing environment deployment engine.The cloud platform in this paper can realize the rapid preparation and deployment of computing environment,the effective management of massive multi-source heterogeneous remote sensing data,and the efficient and automatic processing of remote sensing image change detection.2?In-depth analysis of the FCCRF-CD method,a distributed and parallel method for remote sensing image change detection based on cloud computing is proposed.This paper designs the image reading method in distributed environment based on the characteristics of multi-temporal remote sensing data;combines the K-Means distributed parallel algorithm to optimize the process of unary potential function;improves the calculation process of pairwise potential functions in distributed environments,reduces I / O consumption,and further improves the efficiency of the algorithm.The experimental results show that the distributed parallel method proposed in this paper can significantly improve the detection efficiency on the basis of ensuring accuracy.3?In order to achieve reasonable allocation of cloud platform resources and further improve resource utilization,a multi-objective optimization task scheduling algorithm for remote sensing image change detection based on cuckoo search is proposed.This paper builds an abstract DAG based on the processing flow of change detection algorithms;designs a set of highly scalable virtual machine coding methods based on the state of the virtual machine;establishes a scheduling model with the fastest task completion time and lowest server power consumption as the optimization targets;analyzes several scheduling optimization algorithms,designs the optimization strategy of multi-objective scheduling algorithm based on the implementation mechanism of cuckoo search optimization algorithm.Experimental results show that the multi-objective optimal scheduling algorithm can obtain better scheduling results and can effectively improve the resource utilization of cloud platforms.