复杂散射机制场景的SAR图像认知方法研究

项目名称： 复杂散射机制场景的SAR图像认知方法研究

项目编号： No.61471358

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 无线电电子学、电信技术

项目作者： 邵芸

作者单位： 中国科学院遥感与数字地球研究所

项目金额： 86万元

中文摘要： 针对SAR成像的特殊性和目标/场景的复杂性，选取与人类生存和粮食安全密切相关的城市和稻田场景为研究对象，借鉴人类视觉系统可以高效、快速地处理大量视觉信息的优势，引入认知科学理论与方法，开展复杂散射机制场景的SAR图像认知方法研究。基于微波散射模型、SAR图像模拟、场景三维建模等，研究复杂散射机制场景的SAR图像特征形成机理；借鉴视觉认知机理，基于人类视觉认知习惯与偏好，研究SAR图像视觉优化方法，改善复杂场景SAR图像的呈现方式；结合人类对复杂场景和目标的视觉认知机理，开展机理模型、数据、知识驱动的SAR图像解译模式研究，阐明解译者的认知过程；在散射机理、视觉优化方法和解译模式研究的基础上，针对城市和稻田两类场景，发展适用的多维度信息提取算法。研究成果将有效降低SAR图像应用门槛，推进SAR图像的应用发展。

中文关键词： SAR图像；视觉认知；散射机理；图像解译；参数反演

英文摘要： This project selects the urban area and paddy rice as the research targets to face the challenges of peculiarity of SAR imaging and complexity of target/scene. Urban areas and paddy rice are strongly related to human living environment and food securitty. Human visual system has superior advantage in large amounts of visual information processing and cognition, fast and efficient. Cognitive science theory and method will be utilized in this study. SAR image recognition method applied on the scene with complex scattering mechanisms will be developed. The formation mechanism of SAR imaging will be studied by means of microwave scattering model, SAR image simulation, 3D modeling and formation. Visual optimization method for SAR image will be proposed by utilizing visual perception mechanism, human visual perception, habits and preferences, to improve SAR image understanding and interpretation to the complex scene. SAR image interpretation mode based on the mechanism model, data and knowledge will be developed combining with visual perception mechanism of complex scene and target. Multi dimension information extraction algorithm for two scenarios, urban areas and paddy rice fields, will be developed using scattering mechanism, vision optimization method and interpretation method. The results of this research will effectively reduce the application threshold of the SAR image, and promote the application of SAR image in China.

英文关键词： SAR image;visual cognition;scattering mechanism;image interpretation;parameter retrieval