Skip to main content
ARS Home » Research » Publications at this Location » Publication #255331
Title: Multisensor data fusion of remotely sensed imagery for crop field mapping

Author
item Lan, Yubin
item ZHANG, HUIHUI - Texas A&M University
item Yang, Chenghai
item ZHAO, KAIGUANG - Texas A&M University
item Huang, Yanbo
item Hoffmann, Wesley
item LACEY, RON - Texas A&M University

Submitted to: International Conference on Precision Agriculture Abstracts & Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 7/5/2010
Publication Date: 7/18/2010
Citation: Lan, Y., Zhang, H., Yang, C., Zhao, K., Huang, Y., Hoffmann, W.C., Lacey, R. 2010. Multisensor data fusion of remotely sensed imagery for crop field mapping. International Conference on Precision Agriculture Abstracts & Proceedings. CDROM.

Interpretive Summary: A wide variety of remote sensing data from airborne and satellite platforms is available for site-specific management in agricultural application and production. Aerial image and satellite hyperspectral data were used for mapping crop growth and health status. These results indicate that aerial imagery in conjunction with multisensory data fusion techniques can be a useful tool for mapping crop fields.

Technical Abstract: A wide variety of remote sensing data from airborne and satellite platforms is available for site-specific management in agricultural application and production. Aerial imaging system may offer less expensive and high spatial resolution imagery with Near Infra-Red, Red, Green and Blue spectral wavebands. Space-based hyperspectral sensor such as the Hyperion provides hundreds of spectral bands with low-resolution. Multisensor data fusion provides an effective paradigm for remote sensing applications by synthesizing data from multiple sensors or sources. A high-resolution CIR aerial photo can be integrated with low-resolution hyperspectral images to complement each other for the improved information extraction. Aerial photos can be processed by the state-of-the-art image segmentation algorithms to generate individual objects that often correspond to physically meaningful entities, e.g., a crop land unit growing cotton. Hyperspectral data, commonly with hundreds of spectral bands, can be analyzed using conventional approaches such as spectral indices or possibly more effectively, using advanced data mining models or machine learning algorithms such as support vector machines and Gaussian processes models for accurate retrieval of quantitative information.