Parallel Algorithm For Remote Sensing Big Data In Preventing Re-embedded Digital Watermarking

With the advancement of geo-observation technology,remote sensing big data has been formed because the speed of image acquisition has accelerated,the update cycle of image has shortened,and the amount of data has increased dramatically.However,the illegal acts such as data piracy and infringement have also gradually increased because the application of remote sensing data has been expanding and users are increasing with the advent of the era of remote sensing big data,which seriously jeopardizes national security.Therefore,it is particularly important to take effective technical measures to protect the security of remote sensing big data.Digital watermarking technology,as a frontier technology developed in recent years in the field of information security,can integrate watermark information and data by embedding watermark information such as copyright and user in the carrier data,which can effectively protect the copyright and traceability of data.At present,there are many achievements in digital watermarking technology for remote sensing images and many data rights units use digital watermarking technology to protect the copyright of data.Digital watermarking technology plays an important role in the protection of remote sensing images.However,new problems such as a sharp increase in the amount of data and a dramatic increase in the number of users have come to the fore in the era of remote sensing big data.On the one hand,the existing single-machine watermarking algorithms are difficult to handle remote sensing big data because of the distributed storage.On the other hand,different users repeatedly embed copyright information in the same remote sensing image,resulting in the fact that data copyright ownership is unknown.Therefore,the existing digital watermarking technology cannot meet the copyright protection needs of remote sensing big data.The model and algorithms in preventing re-embedded digital watermarking for remote sensing big data have been studied in this paper,based on the features of remote image and big data.The main research work is as follows:(1)Analyzing the digital watermarking technology may result the problem of the copyright being unknown because of re-embedding of watermark information in the process of sharing remote sensing images.To solve this problem,the concept of watermark identification is proposed,and a preventing re-embedded digital watermarking model based on watermarking pre-detection and authorization mechanism is built,which establishes a theoretical basis for the preventing re-embedded digital watermarking algorithms.(2)The characteristics of image pixel airspace and transform domain coefficients are analyzed.Proposing a digital image watermarking algorithm based on spatial domain-transformation domain.The watermark information is embedded in the transform domain to improve the robustness of the watermark information,while the watermark identifier is embedded in the spatial domain to indicate that the image has embedded the watermark information.The algorithm can not only prevent the watermark information are re-embedded,but also can prevent the watermark information from being re-embedded and correctly extract the watermark information after the remote sensing image suffers from attacks such as clipping.(3)Proposing an anti-repetition digital watermarking algorithm based on transform domain,which integrates watermark identification and watermark information.The algorithm makes full use of the advantages of DCT transform in anti-clipping,anti-compression and other aspects.The watermark identifier and the watermark information are fused to generate a watermark to be embedded in the image DCT coefficients at the same time,and a pre-detection mechanism is added before watermark embedding to prevent the watermark information from being re-embedded.The experimental results show that the algorithm is robust in anti-repetition control and watermark information extraction,and it can resist attacks such as noise,filtering,compression,and cropping.The algorithm is highly efficient and robust.(4)Based on the MapReduce programming framework and the parallel computing methods,proposing a synchronous parallel and a pipeline parallel algorithm for remote sensing big data digital watermarking.Among them,the synchronous parallel algorithm combines the advantages of MapReduce to distribute the computational tasks.The pipeline parallel algorithm optimizes MapReduce from data segmentation and node scheduling after analyzing various stages of MapReduce,which combines the watermarking with MapReduce.The experimental results show that the synchronous parallel watermarking algorithm is more efficient than the standalone serial efficiency in watermark identification pre-detection,watermark embedding and extraction.The pipeline parallel watermarking algorithm is more efficient than the synchronous parallel watermarking algorithm because it optimizes the MapReduce.It is more advantageous when dealing with remote digital image watermarking.