Development of innovative fusion strategies and methods to improve vegetation characterization from multi-sensor remotely sensed data

开发创新的融合策略和方法，以改善多传感器遥感数据的植被特征

• 批准号: RGPIN-2015-06563
• 负责人: Hu, Baoxin
• 金额: $ 1.6万
• 依托单位: York University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2016
• 资助国家: 加拿大
• 起止时间: 2016-01-01 至 2017-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=614223
• 关键词: Development; innovative; fusion; strategies; methods

The long term goal of this proposed research program is to develop innovative integration strategies and fusion methods for the scientific advancement related to the accurate characterization of vegetation canopies from remotely sensed data. It is motivated and driven by technological developments in remote sensing and the demand for advancing science and innovation in vegetation characterization. Rapidly developed remote sensing technologies are making earth observation data widely available in unprecedented volume and detail; and meanwhile recent years have witnessed an increased demand for accurate determination of a growing number of attributes of vegetation canopies using remote sensing for sustainable management of forest resources, environment protection, and precision agriculture. The critical question we currently face is how to effectively utilize these data and intelligently integrate them together to improve vegetation characterization, such as the determination of vegetation types and conditions. Even though data/information fusion is not a new topic in the remote sensing community, the increasing number and heterogeneity of information sources, coupled with the complexity of vegetation canopies leads to an increasing demand for advanced methods  to fully exploit the available technologies, geospatial data, and accumulated knowledge. In this research, I propose innovative integration strategies that are active, intelligent and adaptive, and physics-and knowledge-based, representing a new paradigm in data fusion. These strategies deal with not only methodologies for information combination, but also mechanisms for information selection which actively control and manage the fusion process. With them, the whole scene is not necessarily analyzed in the same way; information and procedures used can be adaptive to local characteristics. The long-term research goal will be achieved through fulfilling the following short-term objectives addressing three important aspects in the characterization of vegetation canopies that are unique in term of algorithm development. (1) Developing intelligent integration approaches to improve individual tree crown delineation. (2) Developing innovative methods to adaptively combine all information obtained from different data sources and improve forest species classification. (3) Developing knowledge-based progressive inversion strategies to improve the retrieval of vegetation parameters using the multi-source remotely sensed data. The success of this program will have a great impact on research and development of information fusion and vegetation characterization. Research results will generate new knowledge and improve our understanding of different sensing technologies and their integrations in vegetation characterization and advance scientific and industrial applications.

这项研究计划的长期目标是开发创新的整合策略和融合方法，以促进与遥感数据准确表征植被冠层相关的科学进步。它的动机和驱动因素是遥感技术的发展以及对推进植被特征描述方面的科学和创新的需求。快速发展的遥感技术使地球观测数据以前所未有的数量和详细程度广泛提供;同时，近年来，人们越来越需要利用遥感准确确定越来越多的植被冠层属性，以实现森林资源的可持续管理、环境保护和精准农业。我们目前面临的关键问题是如何有效地利用这些数据，并将它们智能地整合在一起，以改善植被特征，例如确定植被类型和条件。 尽管数据/信息融合在遥感界并不是一个新的课题，但信息源的数量和异质性不断增加，加上植被冠层的复杂性，导致对先进方法的需求不断增加，以充分利用现有技术、地理空间数据和积累的知识。在这项研究中，我提出了创新的集成策略，是积极的，智能的和自适应的，物理和知识为基础的，代表了一个新的范式，在数据融合。这些策略不仅涉及信息组合的方法，而且还涉及主动控制和管理融合过程的信息选择机制。有了它们，整个场景不一定以相同的方式进行分析;所使用的信息和程序可以适应当地的特点。 长期研究目标将通过实现以下短期目标来实现，这些目标涉及在算法开发方面独特的植被冠层表征的三个重要方面。(1)开发智能集成方法，以改善个体树冠描绘。(2)开发创新方法，自适应地将从不同数据源获得的所有信息联合收割机结合起来，改进森林物种分类。(3)发展以知识为基础的渐进式反演策略，以改善多源遥感数据的植被参数反演。 该项目的成功将对信息融合和植被特征化的研究与发展产生重大影响。研究结果将产生新的知识，提高我们对不同传感技术及其在植被表征中的整合的理解，并推进科学和工业应用。