Research On The Key Technology Of Heterogeneous Sensor Intelligent Surveillance Network

Heterogeneous sensor intelligent surveillance network is a real-time cooperative surveillance system composed of multiple sensor nodes that are deployed in a specific surveillance area,carry various kinds of sensors and possess the ability of sensing,computing and communicating.The system is capable of collecting and processing the information in the surveillance area,and can transmit the information to the control center through wireless network,to achieve the function of remote monitoring.The system has the advantages of rapid deployment,accurate perception,low cost and excellent performance,which has been applied in many fields,such as security guard,battlefield situational awareness,environmental monitoring,traffic management and so on.However,the current research on HSISN focuses on the theoretical analysis,and few studies have been conducted on the practical design of the system.In addition,the large amount of images to be transmitted in the system result in serious data collision and sharp drop in real-time performance.Aiming to solve the above problems,this paper designs the hardware and software of the system based on the goal of designing a practical,intelligent and energy-efficient system.Moreover,to tackle with the problem of giant wireless transmission task of images in the system,this paper proposes an intelligent and energy-efficient image transmission algorithm based on region of interest extraction.The main research contents are as follows:1.Various novel sensors are designed.To improve the detection performance of HSISN sensor nodes,a novel pyroelectric infrared sensor is developed,which adopts two infrared probes,two flat Fresnel lens and the 3rd generation of PIR human pyroelectric infrared detection technology.To reduce the power consumption of the sensor nodes,a novel microwave induction displacement module is designed,which is powered by a pulse power supply with continuously adjustable frequency and duty cycle.The experiment results show that the maximum detection distance of the designed pyroelectric infrared sensor and the microwave induction displacement module is over 40 m,which effectively enhances the detection performance of the sensor nodes and meets the large range monitoring requirement of HSISN.Moreover,the high power module of the sensor node-the microwave induction displacement module saves about 90% of the power consumption and effectively reduces the power consumption compared with the constant power supply,which greatly extends the lifespan of HSISN.2.The prototype of HSISN is developed.According to the design task of HSISN,the overall structure,the working principle and the working mechanism of HSISN are designed in detail.In addition,the main performance metrics of the system are proposed,which provide the technical standards for the subsequent hardware and software design of HSISN nodes.Next,the hardware of other sensors,the MCU,the power module,the communication interface,the isolation protection circuit and the silicone gel heating dew-proof circuit in the nodes are designed.What's more,the low-power design,the anti-interference design,the sealing design and the structure design of HSISN nodes are carried out.Moreover,according to the characteristics of the system and the working mechanism of each node,the embedded program of each node and the communication protocol between nodes are developed.Finally,the PC software for parameters configuration,images display and state monitoring of the prototype is developed by Visual Studio 2015.3.A region of interest extraction algorithm for image wireless transmission is proposed.Firstly,the development of ROI extraction algorithm for images is discussed,and the merits,demerits and applications of various ROI extraction algorithms are analyzed.Secondly,to extract the ROI accurately,robustly,simply and efficiently,the model of the background is established by using ARMA model,and then the parameters of the background model are dynamically updated with the least mean square algorithm to adapt to the complex dynamic background.Finally,each pixel of images is classified into foreground or background.Tests on the CDnet2014 dataset,the Wallflower dataset and the HSISN2019 dataset actually collected show that the proposed algorithm is simple and robust to accurately extract the ROI,and can counteract the dynamic and complex disturbance in the background of images.Moreover,this scheme can achieve the lightweight task of image transmission,reduce the pressure of data wireless transmission in HSISN,and improve the whole lifespan of the system.