|
Hokkaido University Collection of Scholarly and Academic Papers >
Graduate School of Agriculture / Faculty of Agriculture >
Peer-reviewed Journal Articles, etc >
Mapping the spatial distribution of botanical composition and herbage mass in pastures using hyperspectral imaging
| Title: | Mapping the spatial distribution of botanical composition and herbage mass in pastures using hyperspectral imaging |
| Authors: | Suzuki, Yumiko Browse this author | | Okamoto, Hiroshi Browse this author →KAKEN DB | | Takahashi, Makoto Browse this author | | Kataoka, Takashi Browse this author →KAKEN DB | | Shibata, Youichi Browse this author |
| Keywords: | Field map | | herbage mass | | image processing | | plant species | | spectral analysis |
| Issue Date: | Mar-2012 |
| Publisher: | Blackwell Publishing |
| Journal Title: | Grassland Science |
| Volume: | 58 |
| Issue: | 1 |
| Start Page: | 1 |
| End Page: | 7 |
| Publisher DOI: | 10.1111/j.1744-697X.2011.00239.x |
| Abstract: | Quantifying the vegetation heterogeneity in pastures is important to understand the behavior and performance of grazing animals. In this study, we developed a technique for mapping spatial distribution of botanical composition (perennial ryegrass, white clover and other plants) and herbage mass of perennial ryegrass in a grazed pasture using a hyperspectral imaging system, which can glean more detailed spectral information than conventional multispectral imaging that uses standard band frequencies. Plant portions were extracted from hyperspectral image using normalized difference vegetation index threshold and classified into the three plant species or species groups using linear discriminant analysis models. The overall success rate of plant extraction and discrimination between plant species or species groups was 91.6 and 80.3%, respectively. Herbage mass of perennial ryegrass was estimated using partial least squares regression analysis, where the estimated values showed a coefficient of determination of 0.60 against the observed values. In consequence of the models, our results suggested that the difference of leaf cell structure and photosynthesis pigments influenced discrimination between plant species or species group and estimation of the herbage mass. The results demonstrate the potential of the hyperspectral imaging system for creating high spatial resolution maps of vegetation characteristics in pastures. |
| Rights: | The definitive version is available at wileyonlinelibrary.com |
| Type: | article (author version) |
| URI: | http://hdl.handle.net/2115/52104 |
| Appears in Collections: | 農学院・農学研究院 (Graduate School of Agriculture / Faculty of Agriculture) > 雑誌発表論文等 (Peer-reviewed Journal Articles, etc)
|
Submitter: 岡本 博史
|
