Research On Visual Wireless Environment Monitoring System Based On Compressed Sensing

Wireless sensor networks have been widely used in various environmental monitoring situations due to its wide distribution,strong flexibility and convenient layout,which can achieve the timely acquisition of simple data.However,the acquisition and processing of image signal is limited by the processing capacity and energy consumption of sensor nodes.The proposal of compressed sensing theory provides a new direction for image acquisition and processing.Based on the compressed sensing theory,this thesis realizes the image signal acquisition,processing and wireless transmission,which provides a new hardware implementation for the combination of compressed sensing theory and wireless sensor network.In this thesis,firstly,compressed sensing theory and its applications in image processing are briefly introduced,and the common transform of image sparse representation is analyzed in detail.In addition,discrete wavelet transform is selected as image sparse representation base.By analyzing the construction principles and advantages,disadvantages of commonly used measurement matrices,an improved matrix based on some Hadamard measurement matrices is designed,Then,simulation experiments are carried out to verify the correlation with the wavelet transform base,the quality of the reconstructed image and sampling data.The experimental results show that the stability and reconstruction quality of the improved measurement matrix are significantly improved,and the amount of sampling data is less;The implementation principle of OMP algorithm is studied and analyzed.To solve the problem that OMP algorithm must rely on known sparsity and single atom selection in updating support set,selection criteria and regularization processing are used to preprocess atomic selection.Finally,flag bits are added to determine the final selection method of atomic support set.It is verified that the improved algorithm has good reconstruction efficiency and reconstruction quality.According to the sampling characteristics of compressed sensing and the low power consumption of wireless sensor networks,a sensor node based on STM32F407 core board is designed,and some peripheral devices such as OV5640 image acquisition module,n RF24L01 wireless communication module and infrared sensing module are added to realize image acquisition,processing and wireless transmission.At the same time,the star network topology based on n RF24L01 is designed to ensure the stability an reliability of data transmission between nodes.Then,based on the hardware platform,the embedded implementation process of compressed sensing theory is analyzed in detail.Two trigger modes of image acquisition module and drivers of each module are designed.In the sink node,4G module is used to upload serial data to One NET server.Lab VIEW software and MATLAB software are used to jointly develop the upper computer to realize the functions of server data reading,image reconstruction and display.In the part of system verification,the monitoring system is tested in indoor environment and real environment.The experimental results show that the system realizes the organic combination of compressed sensing theory and wireless sensor network.Through a small number of measured values,it can complete the high-quality reconstruction of image signal,and meet the design requirements of visual environmental monitoring.