Proceedings

Find matching any: Reset

Add filter to result:

A Method for Combining Spatial and Hyperspectral Information for Delineation of Homogenous Management Zones

¹Y. Cohen, ¹V. Alchanatis, ²O. Levi, ²S. Cohen

1. ARO
2. BGU

Hyperspectral (HS) remote sensing is a constantly developing field. New remote sensing applications of different fields constantly appear. The possibility of acquisition information about an object without physical contact is spanning new opportunities in many fields and for precision agricultural in particular. These opportunities demand constant improvement and development of new analysis approaches and algorithms, which will exploit the advantages of HS imaging. Though hyperspectral imaging has been transformed in the last 30 years, still most available HS data processing algorithms analyze the data based on the spectral information exclusively and do not treat the data as an image. As for the methods that use spatial and spectral information, most of them use the information serially, as a two-step processing technique. The first processing steps focus on the spectral information, and the second step focus on the spatial one. In this paper we present a classification method which uses both spectral and spatial information simultaneously. A comparison between classification results of different processing approaches is presented. A hyper spectral image of an experimental field of potatoes, where five different levels of fertilizers were applied, was used to test and compare the algorithms’ performance. Three pre-processing models were applied on the hyperspectral image: calculation of a spectral index, principle component analysis and the full HS image. It was found that the principle components and the whole spectra performed better than the spectral index. Furthermore, from the comparison of the classification results, the new approach yielded better classification results than other similar methods reported in the literature.

Keyword: Nitrogen, management zones, beamlets, classification

Y. Cohen V. Alchanatis O. Levi S. Cohen Machine Vision / Multispectral & Hyperspectral Imaging Applications to Precision Agriculture Poster 2012

Download paper