Research On Large-scale Public Remote Sensing Data Acquisition And Scheduling Method In Crowdsourcing Mode

Space remote sensing technology is entering the era of industry application.Landsat,Sentinel and other public welfare satellite remote sensing data play an important role in it,and the user groups of remote sensing data such as individuals,small and medium-sized enterprises,etc.have gradually become larger.In the process of data application,public welfare satellites are generally published on the Internet to global users.Due to the limitations of the service capabilities of data sources and the constraints of data user nodes collection capabilities,largescale public remote sensing data collection has low efficiency and low user collection nodes utilization,which problem still exists.In order to solve the above problems,this paper proposes a large scale public remote sensing data collection and scheduling method under crowdsourcing mode.Firstly,the crowdsourcing acquisition mode architecture is designed.a blockchain trust model is proposed to solve the trust problem between task participants and publishers in the process of crowdsourcing.In order to mobilize the enthusiasm of the majority of task participants,a crowdsourcing incentive model in which both internal and external incentives coexist is proposed.Secondly,the factors that affect the network transmission capacity of crowdsourced nodes are analyzed from the data source,transmission medium and collection terminal.Starting from the collection terminal,some controllable factors are selected to establish the node network transmission capacity prediction model.By improving the inertia weight,learning factor and introducing the precision dynamic adjustment function,an improved PSO-BP algorithm is proposed and compared with the BP algorithm.The results show that the improved PSO-BP algorithm has faster convergence speed and better prediction results.Finally,after analyzing the problem of multi-objective optimization task scheduling,considering the differences in execution speed,execution energy consumption and execution cost of each collection node when executing tasks,a multi-objective task scheduling model based on the task publisher is established,and then separately combining FCFS algorithm,SJF algorithm,MAX?MAX algorithm and PSO algorithm,so an improved multi-objective PSO optimization algorithm is proposed,and multiple sets of simulation comparison experiments are designed.Under the theoretical guidance of the large-scale public remote sensing data collection scheduling method under the crowdsourcing model,the public remote sensing data collection and management system "Crowd4RSData" was designed and built,and the core modules were displayed to show their effects.The two-stage combinatorial verification algorithm named TS-CA is proposed to predict and then make decision,and conducted experiments on a real acquisition system.The experimental results are good,and it also shows that all the research methods mentioned in this paper are indeed effective and feasible,and the public remote sensing data collection and management system built has good application value.