High-speed Reception Management And Preprocessing Of Large-Scale Images

With the continuous development of physics,information science and electronic technology,remote sensing technology and remote sensing imaging level have been greatly improved.As a result,remote sensing images with higher resolution and larger size have been produced.In recent years,the processing of large-scale images has become a research hotspot.In addition,better and faster transmission and management of large-scale images can provide convenience for subsequent image processing.Therefore,this thesis mainly focuses on the study of the transmission and management of large-scale image.In this thesis,the transmission,management and preprocessing of large-scale image are studied.On this basis,a software for high speed reception,management and preprocessing of large-scale image is designed.First of all,this thesis uses modular design ideas to design and develop software according to functional requirements.The designed software is mainly divided into four modules: transmission module,management module,preprocessing module and image display module.In the transmission module,the improved UDP protocol with check code,sending and retransmission is used as the transmission protocol of this design for high-speed transmitson.To solve the problem of large-scale image size,the technology of segmentation and merging is applied on the sender and the receiver respectively.In order to achieve high-speed transmission of the large-scale image,the large-scale image is segmented at the sender and the small sub-images are transmitted in turn.Then,the segmented images are merged at the receiver;In the management module,the My SQL database technology is utilized to create image index database for later retrieval,and the memory management function of the software is realized by process management technology and thread synchronization technology;In the image display module,the adaptive matching method is used to solve the problem of image display adaptability;In the image preprocessing module,the image denoising and registration are realized respectively.In the aspect of image denoising,this thesis analyzes the advantages and disadvantages of the image denoising algorithm based on Wavelet-NSCT.Aiming at the problem of image blur and detail loss caused by this algorithm,the algorithm is improved,and an improved image denoising algorithm based on new threshold and threshold function is proposed.Simulation experiments show that the image denoising algorithms used in this thesis have better denoising effect.In the aspect of the image registration,the registration algorithm based on SIFT feature points is analyzed.Aiming at the problem of feature points redundance and high-dimensional descriptor caused by this algorithm,an improved method of corner detection and modifying feature descriptors is proposed.Simulation experiments show that the improved SIFT image registration algorithm has better image registration accuracy and higher registration speed.