Research Of The Image Processing Method Based On Wireless Sensor Networks

With the complexity of monitoring environment, it is urgent to introduce rich media (i.e.audio, image and video) into environment monitoring activities on the basis of sensor networks, so to perform a fine-grained and accurate environment monitoring. Image Sensor Networks are distributed perception networks with a group of image/video sensor nodes owning the capabilities of computing, storage and communication. As the advanced form of sensor networks, Image sensor networks play an important role in many influential domains such as environment monitoring, security monitoring, health care and so on.However, the sensor nodes's energy, storage and processing capability are limited, so information processing technology based on wireless image sensor networks has caused a great concern. According to the actual requirements of scenes in the safety monitoring system based on energy-constrained wireless image sensor network environment, this paper puts forward an image processing method that can effectively save energy consumption for the image sensor nodes.The main works are as follow:1. The basic conceptions, architecture, features of wireless image sensor networks are analysed, the realated theories of image processing are introduced as well.2. Aiming at the practial requirements that wireless sensor network applying in environments security monitoring, A low power distributed algorthm on image coding and transmission is proposed: the processing method of background subtraction is first employed to detect targets in cluster head node, then a image compression mechanism which combines integer(5,3)wavelet transform based on lifting with SPIHT are used in nodes. It shares the wavelet processing task in each single node and one wavelet decomposition level per cluser. The paper discuss the relationship between several key parameters and energy consuming, such as coding rate, wavelet decomposition level, node numbers, the distance of nodes, hop, PSNR and so on. Simulation results verify the effectivity of this method.3. The energy situation based on event-driven multi-cluster image sensor networks network is analysised on the basis of the distributed image processing scheme. The energy consumption is shown for different levels of wavelet decomposition and image quality requirements. Simulation results show that the proposed algorithm can be effective in reducing the transmission of data, saving energy consumption of nodes.