面向野生动物监测的无线传感器网络图像高效编码与传输方法

Wildlife monitoring based on wireless image sensor networks has a wide application prospect. Under realistic conditions, there are several challenges for wireless image sensor networks including complex sensing environment, limited power consumption of sensors, low reconstructed quality of the interested region of the image, slow and unstable image transmission. The key problems consist of image rectification under complex lighting conditions, image compression and coding, and image transmission. High-efficiency coding and transmission of monitoring images is a critical problem in the actual development of wireless image sensor networks. It is a hot research topic which involves multidisciplinary such as wireless communication, image processing and wildlife protection. This project aims to develop a high-efficiency image coding and transmission approach of wireless image sensor networks for wildlife monitoring. First of all, an image rectification method will be designed for image acquisition in all-weather environment. An adaptive brightness reference value model is established to improve the usability of images obtained under complex illumination conditions. Secondly, an adaptive block-based compressed sensing algorithm will be investigated based on regions of interest in images to reduce the image compression and coding complexity while guarantee a high reconstruction quality of the regions of interest. Furthermore, progressive image transmission scheme based on compressed sensing will be explored to reduce the redundant data and enhance the efficiency and stability in image transmission. By research in these areas, a theoretical innovation and technological breakthroughs for efficient image coding and transmission in wireless sensor networks can be achieved, thereby promoting the application of wireless sensor networks in wildlife monitoring.

无线图像传感器网络在野生动物监测领域有着广泛应用前景，面临感知环境复杂、传感器功耗限制强、图像重点区域重构质量不高、图像传输缓慢且不稳定的挑战，核心问题包括复杂光照条件下图像矫正方法、图像压缩编码方法、图像传输策略等。监测图像的高效编码与传输是限制其推广应用的关键，涉及无线通信、图像处理、野生动物保护等多学科方向交叉，是当前国际研究的热点。本项目力图实现面向野生动物监测的无线传感器网络图像高效编码与传输，研究野外全天候环境下采集图像的矫正方法，建立自适应亮度基准值模型，提高复杂光照条件下采集图像的可用性；研究基于感兴趣区域的图像自适应分块压缩感知算法，保障图像重点区域重构质量，降低图像编码复杂度；寻求基于压缩感知的图像渐进传输方案，减少冗余数据传输，提高传输效率和稳定性。在无线传感器网络图像高效编码与传输理论及关键技术研究上取得突破，推动无线传感器网络在野生动物监测领域的广泛应用。

为了解决无线传感器网络在野生动物监测应用中存在的感知环境复杂、传感器功耗限制强、图像重点区域重构质量不高、图像传输缓慢且不稳定等问题，本项目围绕复杂光照条件下图像矫正方法、图像高效压缩编码以及图像渐进式传输策略等关键问题展开研究。针对复杂光照条件下图像自适应矫正方法，提出一种基于Retinex理论的野生动物监测图像的自适应矫正算法，通过基于改进引导滤波的入射光估计、基于Otus阈值的图像自适应拉伸以及反射光图计算等实现野生动物监测图像的自适应矫正，该算法可以有效克服传统图像增强算法存在的过增强、伪光晕和灰化效应等缺陷，实现不同光照条件下野生动物监测图像的自适应增强；针对野生动物监测图像显著性目标区域提取研究，提出一种基于自适应mean-shift的目标区域提取算法，其基于自适应mean-shift算法，结合颜色空间重构与Hermite滤波器，实现野生动物监测图像目标提取的优化，提高了目标提取的准确率；针对图像高效压缩编码研究，提出一种基于分层感知的图像高效压缩编码算法，在图像显著性目标区域提取的基础上，采用位平面提升和混合编码策略保证图像中重要区域的优先传输，实现了显著性目标区域的无损压缩以及背景区域的有损压缩，保证了图像中重要区域信息的重构质量；针对图像渐进式传输策略研究，在显著性目标检测的基础上，通过对联合信号进行独立编码和联合解码的方法提出一种基于分布式传输机制的图像数据分配策略，实现野生动物监测图像显著性目标区域和背景区域分布式协同传输，实现网络资源的合理利用。本项目在无线传感器网络图像高效压缩编码与传输理论以及关键技术研究上取得突破，为基于无线传感器网络构建野生动物监测系统的推广及应用做出贡献。

内容获取失败，请点击重试