• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Sensors, systems, and next-generation satellites XX >Simulating the directional, spectral and textural properties of a large-scale scene at high resolution using a MODIS BRDF product
【24h】

Simulating the directional, spectral and textural properties of a large-scale scene at high resolution using a MODIS BRDF product

机译：使用MODIS BRDF产品以高分辨率模拟大型场景的方向，光谱和纹理特性

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Many remote sensing applications rely on simulated scenes to perform complex interaction and sensitivity studies that are not possible with real-world scenes. These applications include the development and validation of new and existing algorithms, understanding of the sensor's performance prior to launch, and trade studies to determine ideal sensor configurations. The accuracy of these applications is dependent on the realism of the modeled scenes and sensors. The Digital Image and Remote Sensing Image Generation (DIRSIG) tool has been used extensively to model the complex spectral and spatial texture variation expected in large city-scale scenes and natural biomes. In the past, material properties that were used to represent targets in the simulated scenes were often assumed to be Lambertian in the absence of hand-measured directional data. However, this assumption presents a limitation for new algorithms that need to recognize the anisotropic behavior of targets. We have developed a new method to model and simulate large-scale high-resolution terrestrial scenes by combining bi-directional reflectance distribution function (BRDF) products from Moderate Resolution Imaging Spectroradiometer (MODIS) data, high spatial resolution data, and hyperspectral data. The high spatial resolution data is used to separate materials and add textural variations to the scene, and the directional hemispherical reflectance from the hyperspectral data is used to adjust the magnitude of the MODIS BRDF. In this method, the shape of the BRDF is preserved since it changes very slowly, but its magnitude is varied based on the high resolution texture and hyperspectral data. In addition to the MODIS derived BRDF, target/class specific BRDF values or functions can also be applied to features of specific interest. The purpose of this paper is to discuss the techniques and the methodology used to model a forest region at a high resolution. The simulated scenes using this method for varying view angles show the expected variations in the reflectance due to the BRDF effects of the Harvard forest. The effectiveness of this technique to simulate real sensor data is evaluated by comparing the simulated data with the Landsat 8 Operational Land Image (OLI) data over the Harvard forest. Regions of interest were selected from the simulated and the real data for different targets and their Top-of-Atmospheric (TOA) radiance were compared. After adjusting for scaling correction due to the difference in atmospheric conditions between the simulated and the real data, the TOA radiance is found to agree within 5 ％ in the NIR band and 10 ％ in the visible bands for forest targets under similar illumination conditions. The technique presented in this paper can be extended for other biomes (e.g. desert regions and agricultural regions) by using the appropriate geographic regions. Since the entire scene is constructed in a simulated environment, parameters such as BRDF or its effects can be analyzed for general or target specific algorithm improvements. Also, the modeling and simulation techniques can be used as a baseline for the development and comparison of new sensor designs and to investigate the operational and environmental factors that affects the sensor constellations such as Sentinel and Landsat missions.
机译：许多遥感应用程序都依靠模拟场景来执行复杂的交互作用和灵敏度研究，而这在现实世界的场景中是不可能的。这些应用包括开发和验证新算法和现有算法，在发射之前了解传感器的性能以及进行贸易研究以确定理想的传感器配置。这些应用程序的准确性取决于建模的场景和传感器的真实性。数字图像和遥感图像生成（DIRSIG）工具已被广泛用于模拟大型城市规模的场景和自然生物群落中预期的复杂光谱和空间纹理变化。过去，在没有手工测量的方向数据的情况下，通常被用来表示模拟场景中目标的材料属性是Lambertian。但是，此假设对需要识别目标各向异性行为的新算法提出了限制。通过结合中分辨率成像光谱仪（MODIS）数据，高空间分辨率数据和高光谱数据的双向反射率分布函数（BRDF）产品，我们开发了一种新的方法来建模和模拟大规模高分辨率地面场景。高空间分辨率的数据用于分离材质并向场景添加纹理变化，高光谱数据的定向半球反射率用于调整MODIS BRDF的大小。在此方法中，BRDF的形状由于变化非常缓慢而得以保留，但其大小会根据高分辨率纹理和高光谱数据而变化。除了从MODIS导出的BRDF之外，目标/类特定的BRDF值或函数也可以应用于特定关注的功能。本文的目的是讨论用于以高分辨率对森林区域进行建模的技术和方法。使用此方法改变视角的模拟场景显示了由于哈佛森林的BRDF效应而导致的反射率预期变化。通过将模拟数据与哈佛森林上的Landsat 8可操作土地图像（OLI）数据进行比较，评估了此技术对实际传感器数据的有效性。从模拟中选择了感兴趣的区域，并比较了不同目标的真实数据及其大气顶（TOA）辐射。在模拟和实际数据之间由于大气条件的差异而对比例校正进行调整后，发现在相似光照条件下，森林目标的TOA辐射在NIR波段的可见光范围内在5％以内，在可见波段的可见波段内在10％以内。本文中介绍的技术可以通过使用适当的地理区域扩展到其他生物群落（例如沙漠地区和农业地区）。由于整个场景是在模拟环境中构建的，因此可以分析诸如BRDF或其参数之类的参数，以改进常规或针对特定目标的算法。此外，建模和仿真技术可以用作开发和比较新传感器设计的基准，并可以研究影响传感器星座（例如前哨任务和Landsat任务）的操作和环境因素。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号