Research On Wireless Image Monitoring System Based On Compressed Sensing

Applying Wireless Sensor Network(WSN)in external environment monitoring can change the traditional monitoring system from wired transmission to wireless transmission,making the arrangement of the monitoring system more flexible.However,the problems of low node processing capacity and limited power supply in WSN have not been solved effectively,especially in image processing and transmission.These problems are even more serious.The theory of Compressive Sensing(CS)proposed by research scholars is different from the traditional sampling law.It only needs a small amount of data to accurately reconstruct the original signal.The application of CS theory to sensor nodes can not only reduce the amount of data collected and transmitted by the image,but also solve the problems of limited energy.This paper introduces the CS algorithm into WSN to realize the function of visual image monitoring in the external environment,and provides a new direction for CS theory in hardware implementation and practical engineering applications.The thesis introduces the basic principle and implementation process of CS theory in detail,analyzes three sparse transform methods commonly used in image signals;focuses on several commonly used measurement matrices and their advantages and disadvantages,and designs a structured random matrix that is easy to implement in hardware.The Nuoli measurement matrix verifies the characteristics of RIP,and the improved matrix is compared with commonly used measurement matrices through MATLAB software.The simulation results show that the improved measurement matrix has good observation performance.The matching tracking algorithm and the greedy algorithm are studied.The improved algorithm selects the OMP algorithm as the reconstruction algorithm of the image signal in this paper.Aiming at the existing hardware implementation problems of the original image compression sensing algorithm,an image compression sensing algorithm based on single-layer wavelet transform is designed and implemented by MATLAB software.Analysis and comparison,the experimental results show that the image reconstruction quality of the improved algorithm is improved compared with the original CS algorithm.Aiming at the problems existing in the traditional monitoring system and according to the requirements of low power consumption and flexibility,a new type of sensor node with a Raspberry Pi 3B development board as the main board,OV5647 image sensor,n RF24L01 communication module,and infrared sensor as peripheral devices was designed.The n RF24L01 module sets up a star network topology structure;analyzes the detailed implementation of the image CS algorithm in detail,and writes the CS algorithm and its related control program to the sensor node through the Simulink programming in MATLAB software;the convergent node will receive the The image measurement data was uploaded to Baidu Cloud Disk,and finally formed a complete monitoring system for image acquisition,processing,and transmission.Finally,the system built in this article was tested separately in the laboratory and the actual environment.The experimental results show that the system realizes the combination of compressive sensing theory and wireless sensor network,and can complete the acquisition of image information,the processing of compressive sensing algorithm and the wireless transmission of targets intruding into the monitoring area.When the measurement rate is set to0.6,it can be recovered more accurately in the host computer.