K-nearest neighbor imputation of forest inventory variables in New Hampshire

Andrew J. Lister; Michael Hoppus; Ray L. Czaplewski

Treesearch

K-nearest neighbor imputation of forest inventory variables in New Hampshire

Informally Refereed

Authors:

Andrew J. Lister, Michael Hoppus, Ray L. Czaplewski

Year:

2005

Type:

Scientific Journal

Station:

Northern Research Station

Source:

In: Remote sensing for field users; Proceedings, 10th Biennial USDA Forest Service remote sensing applications conference; 2004 April 5-9; Salt Lake City, UT. Bethesda, MD: American Society for Photogrammetry and Remote Sensing: [unpaginated]. [CD-ROM].

Download (PDF 566 KB)

Download (PDF 566 KB): https://research.fs.usda.gov/download/treesearch/13674.pdf

Abstract

The k-nearest neighbor (kNN) method was used to map stand volume for a mosaic of 4 Landsat scenes covering the state of New Hampshire. Data for gross cubic foot volume and trees per acre were summarized from USDA Forest Service Forest Inventory and Analysis (FIA) plots and used as training for kNN. Six bands of Landsat satellite imagery and various topographic measures were assessed for their use as variables defining the dimensions of the n-dimensional prediction space. Results were generally poor due to the weak correlations between the independent and dependent variables. We discuss the technique and results, and suggest avenues for future research.

Citation

Lister, Andrew; Hoppus, Michael; Czaplewski, Raymond L. 2005. K-nearest neighbor imputation of forest inventory variables in New Hampshire. In: Remote sensing for field users; Proceedings, 10th Biennial USDA Forest Service remote sensing applications conference; 2004 April 5-9; Salt Lake City, UT. Bethesda, MD: American Society for Photogrammetry and Remote Sensing: [unpaginated]. [CD-ROM].