节点文献
面向农业多媒体WSN的彩色图像重构方法研究
Research on Color Image Reconstruction Method for Agricultural Multimedia WSN
【作者】 王强;
【导师】 焦俊;
【作者基本信息】 安徽农业大学 , 农业信息化(专业学位), 2017, 硕士
【摘要】 近年来,随着农业智能化的脚步加快。在农业生产中,搭建无线传感器网络采集单一的环境参数,来监控农业生产环境,已经越来越满足不了农业的生产生活的需要了。同时,随着网络技术和图像处理技术的快速发展,使得面向农业的多媒体无线传感网络应运而生。然而在网络中,传输的数据量过大,不仅导致网络中的采集节点能耗消耗过快,同时会导致网络造成拥塞的问题。因此,为了延长网络采集节点的工作时间,同时降低网络的数据传输量,使用一种合适的压缩和恢复算法已经越来越有着实际的重要意义。本文使用压缩感知将采集的彩色图像进行压缩,并通过重构算法对压缩后的图像进行恢复,从而从理论上减少了数据的传输量。压缩感知(CS),它是新型的信号压缩与恢复理论,分为三个步骤:信号的稀疏表示、测量矩阵和重构算法。通过这三个步骤来对信号进行压缩和重构。本文分析和介绍了CS理论的稀疏表示、测量矩阵和重构算法的数学原理和常用方法,同时介绍了RGB和YUV颜色模型和彩色图像质量判断方法。根据这些基本理论,提出了基于CS理论的RGB和YUV两种颜色模型的彩色图像的压缩和重构算法,将该算法利用于农业彩色图像,并将这些图像分别放在不同稀疏化基、不同测量矩阵、不同重构算法和不同压缩率的对比试验中,进行实验分析。根据实验分析,可得如下结论:基于CS理论的RGB和YUV颜色模型的彩色图像的具有可行性。同时具有高恢复精度。傅立叶稀疏基具有很好的恢复效果,小波稀疏基重构的图像保留了纹理细节。不同的稀疏基和测量矩阵对重构的精度有着很大的影响。稀疏度和压缩率与恢复精度有着极大的相关性。其中,同一压缩率的前提下,稀疏度的加大,重构精度逐渐下降;当在相同稀疏度的情况下,随着压缩率减小,重构精度逐渐提高,然而达到一定的范围时,重构精度就不再增加。
【Abstract】 In recent years,with the fast developing of the agricultural intelligence,building a wireless sensor network to collect a single environmental parameters to monitor the agricultural production environment is unable to adapt the needs of agricultural production and life.At the same time,with the rapid development of network technology and image processing technology,the wireless media sensor network came into being.However,the large data transferred would not only lead to network node energy consumption too fast,but also lead the network congestion in the network.Therefore,it is more and more important and practical significance to use an adaptable compression and recovery algorithm to extend the working time of the network acquisition node and reduce the amount of network data transmission.In this paper,using Compressed Sensing to compressed the collected color image,then using the reconstruction algorithm to recovery the image,which reduces the amount of data transmission.Compression Sensing(CS),which is a new theory of signal compression and recovery,is divided into three steps: the sparse representation of the signal,the measurement matrix and the reconstruction algorithm.The three steps are used to compress and reconstruct the signal.This paper analyzes and introduces the sparse representation of CS theory,the mathematical principles and common methods of measurement matrix and reconstruction algorithm.At the same time,it introduces the RGB and YUV color model and the color image quality judgment method.According to these basic theories,the compression and reconstruction algorithms of color images of RGB and YUV color models based on CS theory are proposed.The agricultural color images experiment is used to verify its feasibility.At the end,these images were placed on different thinning groups,different measurement matrices,different reconstruction algorithms and different compression ratio of the comparative test,the experimental analysis.According to the experimental analysis,the following conclusions can be drawn:RGB and YUV color models of the color image compression perception is feasible.R GB and YUV color models of the color image compression and reconstruction algorithm having high recovery accuracy.Fourier sparse has a good recovery effect,and the wavelet sparse restores the image to preserve the texture details.Different sparse bases and measurement matrices have a significant effect on the accuracy of reconstruction.Sparse degree and compression rate have a great correlation with reconstruction precis ion.In the case of the same degree of sparseness,with the decrease of the compression rate,the reconstruction precision is gradually increased,but when the range of the same degree of sparse degree is increased,the precision of the reconstruction is gradually decreased the precision is no longer increased.