智能控制及其Matlab实现..docVIP

智能控制及其MATLAB实现 题目： 基于BP神经网络的图像压缩实现方法 姓 名： 裴晓鹏 学 号： 2015510298 授课教师： 李国勇 所在院系： 信息工程学院 基于BP神经网络的图像压缩实现方法 摘要 在当今这个信息爆炸的时代，图像传输和共享是一个很重要的环节，但是图像数据 的海量性使其存储和传输成为数字图像处理领域的一大难题。研究发现，一幅图像之所以占据内存大，是因为除了有效信息以外，还有许多冗余信息和相干信息。也正是因为图像中存在着大量的冗余信息和相干信息，使得对图像进行压缩，即去除图像中的冗余信息和相干信息，只保留图像中的有效特征信息变得可行。所谓图像压缩是指用较少的数据量对冗余信息和相干信息进行有效的表征或直接去除。压缩图像为数字图像提供了一种有效的表示方法，同时缓解了系统的存储和传输等压力，在一定程度上能够减少图像存储量和减轻图像传输的负担，也能实现快速传输和实时处理。本文在图像压缩原理及几种常用的图像压缩方法的研究基础上，采用BP神经网络实现图像压缩，能够重构出高质量的图像。在要求存储量较少的同时还追求高质量的重构图像的情况下，这种压缩方法对数据的存储和传输有一定的实践指导作用。 关键词：图像压缩；快速传输；BP神经网络Image compression method based on BP neural network Abstract Today information explodes, the image’s transmission and sharing is an important part but the image data huge amounts make its storage and transport become a big problem in the field of digital image processing. The study found that an image is to occupy memory is big, because in addition to the valid information, there are a lot of redundant information and irrelevant information. Is precisely because there are a lot of redundant information in the image and correlation information, make the image is compressed, which remove the redundant information in the image and correlation information, only keep the effective feature information of images become feasible. Image compression refers to the use of less amount of data to the redundant information and coherent information for effective characterization or remove directly. Compressed image as the representation method of digital image provides an effective, at the same time reduce the pressure on the system of storage and transmission, etc, to a certain extent, can reduce the image storage and reduce the burden of image transmission, also can realize fast transmission and real-time processing. Based on the principle of image compression and several common image compression method based on the research of image compression is realized by using